@article{AdnanSrsicVenticichetal.2020, author = {Adnan, Hassan Sami and Srsic, Amanda and Venticich, Pete Milos and Townend, David M.R.}, title = {Using AI for mental health analysis and prediction in school surveys}, series = {European journal of public health}, volume = {30}, journal = {European journal of public health}, publisher = {Oxford Univ. Press}, address = {Oxford [u.a.]}, issn = {1101-1262}, doi = {10.1093/eurpub/ckaa165.336}, pages = {V125 -- V125}, year = {2020}, abstract = {Background: Childhood and adolescence are critical stages of life for mental health and well-being. Schools are a key setting for mental health promotion and illness prevention. One in five children and adolescents have a mental disorder, about half of mental disorders beginning before the age of 14. Beneficial and explainable artificial intelligence can replace current paper- based and online approaches to school mental health surveys. This can enhance data acquisition, interoperability, data driven analysis, trust and compliance. This paper presents a model for using chatbots for non-obtrusive data collection and supervised machine learning models for data analysis; and discusses ethical considerations pertaining to the use of these models. Methods: For data acquisition, the proposed model uses chatbots which interact with students. The conversation log acts as the source of raw data for the machine learning. Pre-processing of the data is automated by filtering for keywords and phrases. Existing survey results, obtained through current paper-based data collection methods, are evaluated by domain experts (health professionals). These can be used to create a test dataset to validate the machine learning models. Supervised learning can then be deployed to classify specific behaviour and mental health patterns. Results: We present a model that can be used to improve upon current paper-based data collection and manual data analysis methods. An open-source GitHub repository contains necessary tools and components of this model. Privacy is respected through rigorous observance of confidentiality and data protection requirements. Critical reflection on these ethics and law aspects is included in the project. Conclusions: This model strengthens mental health surveillance in schools. The same tools and components could be applied to other public health data. Future extensions of this model could also incorporate unsupervised learning to find clusters and patterns of unknown effects.}, language = {en} } @techreport{Andres2024, type = {Working Paper}, author = {Andres, Maximilian}, title = {Equilibrium selection in infinitely repeated games with communication}, series = {CEPA Discussion Papers}, journal = {CEPA Discussion Papers}, number = {75}, issn = {2628-653X}, doi = {10.25932/publishup-63180}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-631800}, pages = {38}, year = {2024}, abstract = {The present paper proposes a novel approach for equilibrium selection in the infinitely repeated prisoner's dilemma where players can communicate before choosing their strategies. This approach yields a critical discount factor that makes different predictions for cooperation than the usually considered sub-game perfect or risk dominance critical discount factors. In laboratory experiments, we find that our factor is useful for predicting cooperation. For payoff changes where the usually considered factors and our factor make different predictions, the observed cooperation is consistent with the predictions based on our factor.}, language = {en} } @article{AndresBruttelFriedrichsen2022, author = {Andres, Maximilian and Bruttel, Lisa and Friedrichsen, Jana}, title = {How communication makes the difference between a cartel and tacit collusion}, series = {European economic review}, volume = {152}, journal = {European economic review}, publisher = {Elsevier}, address = {Amsterdam}, issn = {0014-2921}, doi = {10.1016/j.euroecorev.2022.104331}, pages = {1 -- 18}, year = {2022}, abstract = {This paper sheds new light on the role of communication for cartel formation. Using machine learning to evaluate free-form chat communication among firms in a laboratory experiment, we identify typical communication patterns for both explicit cartel formation and indirect attempts to collude tacitly. We document that firms are less likely to communicate explicitly about price fixing and more likely to use indirect messages when sanctioning institutions are present. This effect of sanctions on communication reinforces the direct cartel-deterring effect of sanctions as collusion is more difficult to reach and sustain without an explicit agreement. Indirect messages have no, or even a negative, effect on prices.}, language = {en} } @techreport{AndresBruttelFriedrichsen2020, type = {Working Paper}, author = {Andres, Maximilian and Bruttel, Lisa Verena and Friedrichsen, Jana}, title = {Choosing between explicit cartel formation and tacit collusion - An experiment}, series = {CEPA Discussion Papers}, journal = {CEPA Discussion Papers}, number = {19}, issn = {2628-653X}, doi = {10.25932/publishup-47388}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-473885}, pages = {55}, year = {2020}, abstract = {Numerous studies investigate which sanctioning institutions prevent cartel formation but little is known as to how these sanctions work. We contribute to understanding the inner workings of cartels by studying experimentally the effect of sanctioning institutions on firms' communication. Using machine learning to organize the chat communication into topics, we find that firms are significantly less likely to communicate explicitly about price fixing when sanctioning institutions are present. At the same time, average prices are lower when communication is less explicit. A mediation analysis suggests that sanctions are effective in hindering cartel formation not only because they introduce a risk of being fined but also by reducing the prevalence of explicit price communication.}, language = {en} } @techreport{AndresBruttelFriedrichsen2022, type = {Working Paper}, author = {Andres, Maximilian and Bruttel, Lisa Verena and Friedrichsen, Jana}, title = {How communication makes the difference between a cartel and tacit collusion}, series = {CEPA Discussion Papers}, journal = {CEPA Discussion Papers}, issn = {2628-653X}, doi = {10.25932/publishup-56223}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-562234}, pages = {67}, year = {2022}, abstract = {This paper sheds new light on the role of communication for cartel formation. Using machine learning to evaluate free-form chat communication among firms in a laboratory experiment, we identify typical communication patterns for both explicit cartel formation and indirect attempts to collude tacitly. We document that firms are less likely to communicate explicitly about price fixing and more likely to use indirect messages when sanctioning institutions are present. This effect of sanctions on communication reinforces the direct cartel-deterring effect of sanctions as collusion is more difficult to reach and sustain without an explicit agreement. Indirect messages have no, or even a negative, effect on prices.}, language = {en} } @article{Ayzel2021, author = {Ayzel, Georgy}, title = {Deep neural networks in hydrology}, series = {Vestnik of Saint Petersburg University. Earth Sciences}, volume = {66}, journal = {Vestnik of Saint Petersburg University. Earth Sciences}, number = {1}, publisher = {Univ. Press}, address = {St. Petersburg}, issn = {2541-9668}, doi = {10.21638/spbu07.2021.101}, pages = {5 -- 18}, year = {2021}, abstract = {For around a decade, deep learning - the sub-field of machine learning that refers to artificial neural networks comprised of many computational layers - modifies the landscape of statistical model development in many research areas, such as image classification, machine translation, and speech recognition. Geoscientific disciplines in general and the field of hydrology in particular, also do not stand aside from this movement. Recently, the proliferation of modern deep learning-based techniques and methods has been actively gaining popularity for solving a wide range of hydrological problems: modeling and forecasting of river runoff, hydrological model parameters regionalization, assessment of available water resources. identification of the main drivers of the recent change in water balance components. This growing popularity of deep neural networks is primarily due to their high universality and efficiency. The presented qualities, together with the rapidly growing amount of accumulated environmental information, as well as increasing availability of computing facilities and resources, allow us to speak about deep neural networks as a new generation of mathematical models designed to, if not to replace existing solutions, but significantly enrich the field of geophysical processes modeling. This paper provides a brief overview of the current state of the field of development and application of deep neural networks in hydrology. Also in the following study, the qualitative long-term forecast regarding the development of deep learning technology for managing the corresponding hydrological modeling challenges is provided based on the use of "Gartner Hype Curve", which in the general details describes a life cycle of modern technologies.}, language = {en} } @misc{AyzelIzhitskiy2019, author = {Ayzel, Georgy and Izhitskiy, Alexander}, title = {Climate change impact assessment on freshwater inflow into the Small Aral Sea}, series = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {1071}, issn = {1866-8372}, doi = {10.25932/publishup-47279}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-472794}, pages = {21}, year = {2019}, abstract = {During the last few decades, the rapid separation of the Small Aral Sea from the isolated basin has changed its hydrological and ecological conditions tremendously. In the present study, we developed and validated the hybrid model for the Syr Darya River basin based on a combination of state-of-the-art hydrological and machine learning models. Climate change impact on freshwater inflow into the Small Aral Sea for the projection period 2007-2099 has been quantified based on the developed hybrid model and bias corrected and downscaled meteorological projections simulated by four General Circulation Models (GCM) for each of three Representative Concentration Pathway scenarios (RCP). The developed hybrid model reliably simulates freshwater inflow for the historical period with a Nash-Sutcliffe efficiency of 0.72 and a Kling-Gupta efficiency of 0.77. Results of the climate change impact assessment showed that the freshwater inflow projections produced by different GCMs are misleading by providing contradictory results for the projection period. However, we identified that the relative runoff changes are expected to be more pronounced in the case of more aggressive RCP scenarios. The simulated projections of freshwater inflow provide a basis for further assessment of climate change impacts on hydrological and ecological conditions of the Small Aral Sea in the 21st Century.}, language = {en} } @article{AyzelIzhitskiy2019, author = {Ayzel, Georgy and Izhitskiy, Alexander}, title = {Climate Change Impact Assessment on Freshwater Inflow into the Small Aral Sea}, series = {Water}, volume = {11}, journal = {Water}, number = {11}, publisher = {MDPI}, address = {Basel}, issn = {2073-4441}, doi = {10.3390/w11112377}, pages = {19}, year = {2019}, abstract = {During the last few decades, the rapid separation of the Small Aral Sea from the isolated basin has changed its hydrological and ecological conditions tremendously. In the present study, we developed and validated the hybrid model for the Syr Darya River basin based on a combination of state-of-the-art hydrological and machine learning models. Climate change impact on freshwater inflow into the Small Aral Sea for the projection period 2007-2099 has been quantified based on the developed hybrid model and bias corrected and downscaled meteorological projections simulated by four General Circulation Models (GCM) for each of three Representative Concentration Pathway scenarios (RCP). The developed hybrid model reliably simulates freshwater inflow for the historical period with a Nash-Sutcliffe efficiency of 0.72 and a Kling-Gupta efficiency of 0.77. Results of the climate change impact assessment showed that the freshwater inflow projections produced by different GCMs are misleading by providing contradictory results for the projection period. However, we identified that the relative runoff changes are expected to be more pronounced in the case of more aggressive RCP scenarios. The simulated projections of freshwater inflow provide a basis for further assessment of climate change impacts on hydrological and ecological conditions of the Small Aral Sea in the 21st Century.}, language = {en} } @article{BornhorstNustedeFudickar2019, author = {Bornhorst, Julia and Nustede, Eike Jannik and Fudickar, Sebastian}, title = {Mass Surveilance of C. elegans-Smartphone-Based DIY Microscope and Machine-Learning-Based Approach for Worm Detection}, series = {Sensors}, volume = {19}, journal = {Sensors}, number = {6}, publisher = {MDPI}, address = {Basel}, issn = {1424-8220}, doi = {10.3390/s19061468}, pages = {14}, year = {2019}, abstract = {The nematode Caenorhabditis elegans (C. elegans) is often used as an alternative animal model due to several advantages such as morphological changes that can be seen directly under a microscope. Limitations of the model include the usage of expensive and cumbersome microscopes, and restrictions of the comprehensive use of C. elegans for toxicological trials. With the general applicability of the detection of C. elegans from microscope images via machine learning, as well as of smartphone-based microscopes, this article investigates the suitability of smartphone-based microscopy to detect C. elegans in a complete Petri dish. Thereby, the article introduces a smartphone-based microscope (including optics, lighting, and housing) for monitoring C. elegans and the corresponding classification via a trained Histogram of Oriented Gradients (HOG) feature-based Support Vector Machine for the automatic detection of C. elegans. Evaluation showed classification sensitivity of 0.90 and specificity of 0.85, and thereby confirms the general practicability of the chosen approach.}, language = {en} } @article{BrandesSicksBerger2021, author = {Brandes, Stefanie and Sicks, Florian and Berger, Anne}, title = {Behaviour classification on giraffes (Giraffa camelopardalis) using machine learning algorithms on triaxial acceleration data of two commonly used GPS devices and its possible application for their management and conservation}, series = {Sensors}, volume = {21}, journal = {Sensors}, number = {6}, publisher = {MDPI}, address = {Basel}, issn = {1424-8220}, doi = {10.3390/s21062229}, pages = {22}, year = {2021}, abstract = {Averting today's loss of biodiversity and ecosystem services can be achieved through conservation efforts, especially of keystone species. Giraffes (Giraffa camelopardalis) play an important role in sustaining Africa's ecosystems, but are 'vulnerable' according to the IUCN Red List since 2016. Monitoring an animal's behavior in the wild helps to develop and assess their conservation management. One mechanism for remote tracking of wildlife behavior is to attach accelerometers to animals to record their body movement. We tested two different commercially available high-resolution accelerometers, e-obs and Africa Wildlife Tracking (AWT), attached to the top of the heads of three captive giraffes and analyzed the accuracy of automatic behavior classifications, focused on the Random Forests algorithm. For both accelerometers, behaviors of lower variety in head and neck movements could be better predicted (i.e., feeding above eye level, mean prediction accuracy e-obs/AWT: 97.6\%/99.7\%; drinking: 96.7\%/97.0\%) than those with a higher variety of body postures (such as standing: 90.7-91.0\%/75.2-76.7\%; rumination: 89.6-91.6\%/53.5-86.5\%). Nonetheless both devices come with limitations and especially the AWT needs technological adaptations before applying it on animals in the wild. Nevertheless, looking at the prediction results, both are promising accelerometers for behavioral classification of giraffes. Therefore, these devices when applied to free-ranging animals, in combination with GPS tracking, can contribute greatly to the conservation of giraffes.}, language = {en} } @phdthesis{Brill2022, author = {Brill, Fabio Alexander}, title = {Applications of machine learning and open geospatial data in flood risk modelling}, doi = {10.25932/publishup-55594}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-555943}, school = {Universit{\"a}t Potsdam}, pages = {xix, 124}, year = {2022}, abstract = {Der technologische Fortschritt erlaubt es, zunehmend komplexe Vorhersagemodelle auf Basis immer gr{\"o}ßerer Datens{\"a}tze zu produzieren. F{\"u}r das Risikomanagement von Naturgefahren sind eine Vielzahl von Modellen als Entscheidungsgrundlage notwendig, z.B. in der Auswertung von Beobachtungsdaten, f{\"u}r die Vorhersage von Gefahrenszenarien, oder zur statistischen Absch{\"a}tzung der zu erwartenden Sch{\"a}den. Es stellt sich also die Frage, inwiefern moderne Modellierungsans{\"a}tze wie das maschinelle Lernen oder Data-Mining in diesem Themenbereich sinnvoll eingesetzt werden k{\"o}nnen. Zus{\"a}tzlich ist im Hinblick auf die Datenverf{\"u}gbarkeit und -zug{\"a}nglichkeit ein Trend zur {\"O}ffnung (open data) zu beobachten. Thema dieser Arbeit ist daher, die M{\"o}glichkeiten und Grenzen des maschinellen Lernens und frei verf{\"u}gbarer Geodaten auf dem Gebiet der Hochwasserrisikomodellierung im weiteren Sinne zu untersuchen. Da dieses {\"u}bergeordnete Thema sehr breit ist, werden einzelne relevante Aspekte herausgearbeitet und detailliert betrachtet. Eine prominente Datenquelle im Bereich Hochwasser ist die satellitenbasierte Kartierung von {\"U}berflutungsfl{\"a}chen, die z.B. {\"u}ber den Copernicus Service der Europ{\"a}ischen Union frei zur Verf{\"u}gung gestellt werden. Große Hoffnungen werden in der wissenschaftlichen Literatur in diese Produkte gesetzt, sowohl f{\"u}r die akute Unterst{\"u}tzung der Einsatzkr{\"a}fte im Katastrophenfall, als auch in der Modellierung mittels hydrodynamischer Modelle oder zur Schadensabsch{\"a}tzung. Daher wurde ein Fokus in dieser Arbeit auf die Untersuchung dieser Flutmasken gelegt. Aus der Beobachtung, dass die Qualit{\"a}t dieser Produkte in bewaldeten und urbanen Gebieten unzureichend ist, wurde ein Verfahren zur nachtr{\"a}glichenVerbesserung mittels maschinellem Lernen entwickelt. Das Verfahren basiert auf einem Klassifikationsalgorithmus der nur Trainingsdaten von einer vorherzusagenden Klasse ben{\"o}tigt, im konkreten Fall also Daten von {\"U}berflutungsfl{\"a}chen, nicht jedoch von der negativen Klasse (trockene Gebiete). Die Anwendung f{\"u}r Hurricane Harvey in Houston zeigt großes Potenzial der Methode, abh{\"a}ngig von der Qualit{\"a}t der urspr{\"u}nglichen Flutmaske. Anschließend wird anhand einer prozessbasierten Modellkette untersucht, welchen Einfluss implementierte physikalische Prozessdetails auf das vorhergesagte statistische Risiko haben. Es wird anschaulich gezeigt, was eine Risikostudie basierend auf etablierten Modellen leisten kann. Solche Modellketten sind allerdings bereits f{\"u}r Flusshochwasser sehr komplex, und f{\"u}r zusammengesetzte oder kaskadierende Ereignisse mit Starkregen, Sturzfluten, und weiteren Prozessen, kaum vorhanden. Im vierten Kapitel dieser Arbeit wird daher getestet, ob maschinelles Lernen auf Basis von vollst{\"a}ndigen Schadensdaten einen direkteren Weg zur Schadensmodellierung erm{\"o}glicht, der die explizite Konzeption einer solchen Modellkette umgeht. Dazu wird ein staatlich erhobener Datensatz der gesch{\"a}digten Geb{\"a}ude w{\"a}hrend des schweren El Ni{\~n}o Ereignisses 2017 in Peru verwendet. In diesem Kontext werden auch die M{\"o}glichkeiten des Data-Mining zur Extraktion von Prozessverst{\"a}ndnis ausgelotet. Es kann gezeigt werden, dass diverse frei verf{\"u}gbare Geodaten n{\"u}tzliche Informationen f{\"u}r die Gefahren- und Schadensmodellierung von komplexen Flutereignissen liefern, z.B. satellitenbasierte Regenmessungen, topographische und hydrographische Information, kartierte Siedlungsfl{\"a}chen, sowie Indikatoren aus Spektraldaten. Zudem zeigen sich Erkenntnisse zu den Sch{\"a}digungsprozessen, die im Wesentlichen mit den vorherigen Erwartungen in Einklang stehen. Die maximale Regenintensit{\"a}t wirkt beispielsweise in St{\"a}dten und steilen Schluchten st{\"a}rker sch{\"a}digend, w{\"a}hrend die Niederschlagssumme in tiefliegenden Flussgebieten und bewaldeten Regionen als aussagekr{\"a}ftiger befunden wurde. L{\"a}ndliche Gebiete in Peru weisen in der pr{\"a}sentierten Studie eine h{\"o}here Vulnerabilit{\"a}t als die Stadtgebiete auf. Jedoch werden auch die grunds{\"a}tzlichen Grenzen der Methodik und die Abh{\"a}ngigkeit von spezifischen Datens{\"a}tzen and Algorithmen offenkundig. In der {\"u}bergreifenden Diskussion werden schließlich die verschiedenen Methoden - prozessbasierte Modellierung, pr{\"a}diktives maschinelles Lernen, und Data-Mining - mit Blick auf die Gesamtfragestellungen evaluiert. Im Bereich der Gefahrenbeobachtung scheint eine Fokussierung auf neue Algorithmen sinnvoll. Im Bereich der Gefahrenmodellierung, insbesondere f{\"u}r Flusshochwasser, wird eher die Verbesserung von physikalischen Modellen, oder die Integration von prozessbasierten und statistischen Verfahren angeraten. In der Schadensmodellierung fehlen nach wie vor die großen repr{\"a}sentativen Datens{\"a}tze, die f{\"u}r eine breite Anwendung von maschinellem Lernen Voraussetzung ist. Daher ist die Verbesserung der Datengrundlage im Bereich der Sch{\"a}den derzeit als wichtiger einzustufen als die Auswahl der Algorithmen.}, language = {en} } @article{CeulemansGuillGaedke2021, author = {Ceulemans, Ruben and Guill, Christian and Gaedke, Ursula}, title = {Top predators govern multitrophic diversity effects in tritrophic food webs}, series = {Ecology : a publication of the Ecological Society of America}, volume = {102}, journal = {Ecology : a publication of the Ecological Society of America}, number = {7}, publisher = {Wiley}, address = {Hoboken}, issn = {0012-9658}, doi = {10.1002/ecy.3379}, pages = {16}, year = {2021}, abstract = {It is well known that functional diversity strongly affects ecosystem functioning. However, even in rather simple model communities consisting of only two or, at best, three trophic levels, the relationship between multitrophic functional diversity and ecosystem functioning appears difficult to generalize, because of its high contextuality. In this study, we considered several differently structured tritrophic food webs, in which the amount of functional diversity was varied independently on each trophic level. To achieve generalizable results, largely independent of parametrization, we examined the outcomes of 128,000 parameter combinations sampled from ecologically plausible intervals, with each tested for 200 randomly sampled initial conditions. Analysis of our data was done by training a random forest model. This method enables the identification of complex patterns in the data through partial dependence graphs, and the comparison of the relative influence of model parameters, including the degree of diversity, on food-web properties. We found that bottom-up and top-down effects cascade simultaneously throughout the food web, intimately linking the effects of functional diversity of any trophic level to the amount of diversity of other trophic levels, which may explain the difficulty in unifying results from previous studies. Strikingly, only with high diversity throughout the whole food web, different interactions synergize to ensure efficient exploitation of the available nutrients and efficient biomass transfer to higher trophic levels, ultimately leading to a high biomass and production on the top level. The temporal variation of biomass showed a more complex pattern with increasing multitrophic diversity: while the system initially became less variable, eventually the temporal variation rose again because of the increasingly complex dynamical patterns. Importantly, top predator diversity and food-web parameters affecting the top trophic level were of highest importance to determine the biomass and temporal variability of any trophic level. Overall, our study reveals that the mechanisms by which diversity influences ecosystem functioning are affected by every part of the food web, hampering the extrapolation of insights from simple monotrophic or bitrophic systems to complex natural food webs.}, language = {en} } @phdthesis{Chen2023, author = {Chen, Junchao}, title = {A self-adaptive resilient method for implementing and managing the high-reliability processing system}, doi = {10.25932/publishup-58313}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-583139}, school = {Universit{\"a}t Potsdam}, pages = {XXIII, 167}, year = {2023}, abstract = {As a result of CMOS scaling, radiation-induced Single-Event Effects (SEEs) in electronic circuits became a critical reliability issue for modern Integrated Circuits (ICs) operating under harsh radiation conditions. SEEs can be triggered in combinational or sequential logic by the impact of high-energy particles, leading to destructive or non-destructive faults, resulting in data corruption or even system failure. Typically, the SEE mitigation methods are deployed statically in processing architectures based on the worst-case radiation conditions, which is most of the time unnecessary and results in a resource overhead. Moreover, the space radiation conditions are dynamically changing, especially during Solar Particle Events (SPEs). The intensity of space radiation can differ over five orders of magnitude within a few hours or days, resulting in several orders of magnitude fault probability variation in ICs during SPEs. This thesis introduces a comprehensive approach for designing a self-adaptive fault resilient multiprocessing system to overcome the static mitigation overhead issue. This work mainly addresses the following topics: (1) Design of on-chip radiation particle monitor for real-time radiation environment detection, (2) Investigation of space environment predictor, as support for solar particle events forecast, (3) Dynamic mode configuration in the resilient multiprocessing system. Therefore, according to detected and predicted in-flight space radiation conditions, the target system can be configured to use no mitigation or low-overhead mitigation during non-critical periods of time. The redundant resources can be used to improve system performance or save power. On the other hand, during increased radiation activity periods, such as SPEs, the mitigation methods can be dynamically configured appropriately depending on the real-time space radiation environment, resulting in higher system reliability. Thus, a dynamic trade-off in the target system between reliability, performance and power consumption in real-time can be achieved. All results of this work are evaluated in a highly reliable quad-core multiprocessing system that allows the self-adaptive setting of optimal radiation mitigation mechanisms during run-time. Proposed methods can serve as a basis for establishing a comprehensive self-adaptive resilient system design process. Successful implementation of the proposed design in the quad-core multiprocessor shows its application perspective also in the other designs.}, language = {en} } @article{ChenLangeAndjelkovicetal.2022, author = {Chen, Junchao and Lange, Thomas and Andjelkovic, Marko and Simevski, Aleksandar and Lu, Li and Krstic, Milos}, title = {Solar particle event and single event upset prediction from SRAM-based monitor and supervised machine learning}, series = {IEEE transactions on emerging topics in computing / IEEE Computer Society, Institute of Electrical and Electronics Engineers}, volume = {10}, journal = {IEEE transactions on emerging topics in computing / IEEE Computer Society, Institute of Electrical and Electronics Engineers}, number = {2}, publisher = {Institute of Electrical and Electronics Engineers}, address = {[New York, NY]}, issn = {2168-6750}, doi = {10.1109/TETC.2022.3147376}, pages = {564 -- 580}, year = {2022}, abstract = {The intensity of cosmic radiation may differ over five orders of magnitude within a few hours or days during the Solar Particle Events (SPEs), thus increasing for several orders of magnitude the probability of Single Event Upsets (SEUs) in space-borne electronic systems. Therefore, it is vital to enable the early detection of the SEU rate changes in order to ensure timely activation of dynamic radiation hardening measures. In this paper, an embedded approach for the prediction of SPEs and SRAM SEU rate is presented. The proposed solution combines the real-time SRAM-based SEU monitor, the offline-trained machine learning model and online learning algorithm for the prediction. With respect to the state-of-the-art, our solution brings the following benefits: (1) Use of existing on-chip data storage SRAM as a particle detector, thus minimizing the hardware and power overhead, (2) Prediction of SRAM SEU rate one hour in advance, with the fine-grained hourly tracking of SEU variations during SPEs as well as under normal conditions, (3) Online optimization of the prediction model for enhancing the prediction accuracy during run-time, (4) Negligible cost of hardware accelerator design for the implementation of selected machine learning model and online learning algorithm. The proposed design is intended for a highly dependable and self-adaptive multiprocessing system employed in space applications, allowing to trigger the radiation mitigation mechanisms before the onset of high radiation levels.}, language = {en} } @article{CopeBaukmannKlingeretal.2021, author = {Cope, Justin L. and Baukmann, Hannes A. and Klinger, J{\"o}rn E. and Ravarani, Charles N. J. and B{\"o}ttinger, Erwin and Konigorski, Stefan and Schmidt, Marco F.}, title = {Interaction-based feature selection algorithm outperforms polygenic risk score in predicting Parkinson's Disease status}, series = {Frontiers in genetics}, volume = {12}, journal = {Frontiers in genetics}, publisher = {Frontiers Media}, address = {Lausanne}, issn = {1664-8021}, doi = {10.3389/fgene.2021.744557}, pages = {9}, year = {2021}, abstract = {Polygenic risk scores (PRS) aggregating results from genome-wide association studies are the state of the art in the prediction of susceptibility to complex traits or diseases, yet their predictive performance is limited for various reasons, not least of which is their failure to incorporate the effects of gene-gene interactions. Novel machine learning algorithms that use large amounts of data promise to find gene-gene interactions in order to build models with better predictive performance than PRS. Here, we present a data preprocessing step by using data-mining of contextual information to reduce the number of features, enabling machine learning algorithms to identify gene-gene interactions. We applied our approach to the Parkinson's Progression Markers Initiative (PPMI) dataset, an observational clinical study of 471 genotyped subjects (368 cases and 152 controls). With an AUC of 0.85 (95\% CI = [0.72; 0.96]), the interaction-based prediction model outperforms the PRS (AUC of 0.58 (95\% CI = [0.42; 0.81])). Furthermore, feature importance analysis of the model provided insights into the mechanism of Parkinson's disease. For instance, the model revealed an interaction of previously described drug target candidate genes TMEM175 and GAPDHP25. These results demonstrate that interaction-based machine learning models can improve genetic prediction models and might provide an answer to the missing heritability problem.}, language = {en} } @phdthesis{DokhtDolatabadiEsfahani2022, author = {Dokht Dolatabadi Esfahani, Reza}, title = {Time-dependent monitoring of near-surface and ground motion modelling: developing new data processing approaches based on Music Information Retrieval (MIR) strategies}, doi = {10.25932/publishup-56767}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-567671}, school = {Universit{\"a}t Potsdam}, pages = {xiv, 107}, year = {2022}, abstract = {Seismology, like many scientific fields, e.g., music information retrieval and speech signal pro- cessing, is experiencing exponential growth in the amount of data acquired by modern seismo- logical networks. In this thesis, I take advantage of the opportunities offered by "big data" and by the methods developed in the areas of music information retrieval and machine learning to predict better the ground motion generated by earthquakes and to study the properties of the surface layers of the Earth. In order to better predict seismic ground motions, I propose two approaches based on unsupervised deep learning methods, an autoencoder network and Generative Adversarial Networks. The autoencoder technique explores a massive amount of ground motion data, evaluates the required parameters, and generates synthetic ground motion data in the Fourier amplitude spectra (FAS) domain. This method is tested on two synthetic datasets and one real dataset. The application on the real dataset shows that the substantial information contained within the FAS data can be encoded to a four to the five-dimensional manifold. Consequently, only a few independent parameters are required for efficient ground motion prediction. I also propose a method based on Conditional Generative Adversarial Networks (CGAN) for simulating ground motion records in the time-frequency and time domains. CGAN generates the time-frequency domains based on the parameters: magnitude, distance, and shear wave velocities to 30 m depth (VS30). After generating the amplitude of the time-frequency domains using the CGAN model, instead of classical conventional methods that assume the amplitude spectra with a random phase spectrum, the phase of the time-frequency domains is recovered by minimizing the observed and reconstructed spectrograms. In the second part of this dissertation, I propose two methods for the monitoring and characterization of near-surface materials and site effect analyses. I implement an autocorrelation function and an interferometry method to monitor the velocity changes of near-surface materials resulting from the Kumamoto earthquake sequence (Japan, 2016). The observed seismic velocity changes during the strong shaking are due to the non-linear response of the near-surface materials. The results show that the velocity changes lasted for about two months after the Kumamoto mainshock. Furthermore, I used the velocity changes to evaluate the in-situ strain-stress relationship. I also propose a method for assessing the site proxy "VS30" using non-invasive analysis. In the proposed method, a dispersion curve of surface waves is inverted to estimate the shear wave velocity of the subsurface. This method is based on the Dix-like linear operators, which relate the shear wave velocity to the phase velocity. The proposed method is fast, efficient, and stable. All of the methods presented in this work can be used for processing "big data" in seismology and for the analysis of weak and strong ground motion data, to predict ground shaking, and to analyze site responses by considering potential time dependencies and nonlinearities.}, language = {en} } @article{Doellner2020, author = {D{\"o}llner, J{\"u}rgen Roland Friedrich}, title = {Geospatial artificial intelligence}, series = {Journal of photogrammetry, remote sensing and geoinformation science : PFG : Photogrammetrie, Fernerkundung, Geoinformation}, volume = {88}, journal = {Journal of photogrammetry, remote sensing and geoinformation science : PFG : Photogrammetrie, Fernerkundung, Geoinformation}, number = {1}, publisher = {Springer International Publishing}, address = {Cham}, issn = {2512-2789}, doi = {10.1007/s41064-020-00102-3}, pages = {15 -- 24}, year = {2020}, abstract = {Artificial intelligence (AI) is changing fundamentally the way how IT solutions are implemented and operated across all application domains, including the geospatial domain. This contribution outlines AI-based techniques for 3D point clouds and geospatial digital twins as generic components of geospatial AI. First, we briefly reflect on the term "AI" and outline technology developments needed to apply AI to IT solutions, seen from a software engineering perspective. Next, we characterize 3D point clouds as key category of geodata and their role for creating the basis for geospatial digital twins; we explain the feasibility of machine learning (ML) and deep learning (DL) approaches for 3D point clouds. In particular, we argue that 3D point clouds can be seen as a corpus with similar properties as natural language corpora and formulate a "Naturalness Hypothesis" for 3D point clouds. In the main part, we introduce a workflow for interpreting 3D point clouds based on ML/DL approaches that derive domain-specific and application-specific semantics for 3D point clouds without having to create explicit spatial 3D models or explicit rule sets. Finally, examples are shown how ML/DL enables us to efficiently build and maintain base data for geospatial digital twins such as virtual 3D city models, indoor models, or building information models.}, language = {en} } @article{EbersHochRosenkranzetal.2021, author = {Ebers, Martin and Hoch, Veronica R. S. and Rosenkranz, Frank and Ruschemeier, Hannah and Steinr{\"o}tter, Bj{\"o}rn}, title = {The European Commission's proposal for an Artificial Intelligence Act}, series = {J : multidisciplinary scientific journal}, volume = {4}, journal = {J : multidisciplinary scientific journal}, number = {4}, publisher = {MDPI}, address = {Basel}, issn = {2571-8800}, doi = {10.3390/j4040043}, pages = {589 -- 603}, year = {2021}, abstract = {On 21 April 2021, the European Commission presented its long-awaited proposal for a Regulation "laying down harmonized rules on Artificial Intelligence", the so-called "Artificial Intelligence Act" (AIA). This article takes a critical look at the proposed regulation. After an introduction (1), the paper analyzes the unclear preemptive effect of the AIA and EU competences (2), the scope of application (3), the prohibited uses of Artificial Intelligence (AI) (4), the provisions on high-risk AI systems (5), the obligations of providers and users (6), the requirements for AI systems with limited risks (7), the enforcement system (8), the relationship of the AIA with the existing legal framework (9), and the regulatory gaps (10). The last section draws some final conclusions (11).}, language = {en} } @phdthesis{Elsaid2022, author = {Elsaid, Mohamed Esameldin Mohamed}, title = {Virtual machines live migration cost modeling and prediction}, doi = {10.25932/publishup-54001}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-540013}, school = {Universit{\"a}t Potsdam}, pages = {xiv, 107}, year = {2022}, abstract = {Dynamic resource management is an essential requirement for private and public cloud computing environments. With dynamic resource management, the physical resources assignment to the cloud virtual resources depends on the actual need of the applications or the running services, which enhances the cloud physical resources utilization and reduces the offered services cost. In addition, the virtual resources can be moved across different physical resources in the cloud environment without an obvious impact on the running applications or services production. This means that the availability of the running services and applications in the cloud is independent on the hardware resources including the servers, switches and storage failures. This increases the reliability of using cloud services compared to the classical data-centers environments. In this thesis we briefly discuss the dynamic resource management topic and then deeply focus on live migration as the definition of the compute resource dynamic management. Live migration is a commonly used and an essential feature in cloud and virtual data-centers environments. Cloud computing load balance, power saving and fault tolerance features are all dependent on live migration to optimize the virtual and physical resources usage. As we will discuss in this thesis, live migration shows many benefits to cloud and virtual data-centers environments, however the cost of live migration can not be ignored. Live migration cost includes the migration time, downtime, network overhead, power consumption increases and CPU overhead. IT admins run virtual machines live migrations without an idea about the migration cost. So, resources bottlenecks, higher migration cost and migration failures might happen. The first problem that we discuss in this thesis is how to model the cost of the virtual machines live migration. Secondly, we investigate how to make use of machine learning techniques to help the cloud admins getting an estimation of this cost before initiating the migration for one of multiple virtual machines. Also, we discuss the optimal timing for a specific virtual machine before live migration to another server. Finally, we propose practical solutions that can be used by the cloud admins to be integrated with the cloud administration portals to answer the raised research questions above. Our research methodology to achieve the project objectives is to propose empirical models based on using VMware test-beds with different benchmarks tools. Then we make use of the machine learning techniques to propose a prediction approach for virtual machines live migration cost. Timing optimization for live migration is also proposed in this thesis based on using the cost prediction and data-centers network utilization prediction. Live migration with persistent memory clusters is also discussed at the end of the thesis. The cost prediction and timing optimization techniques proposed in this thesis could be practically integrated with VMware vSphere cluster portal such that the IT admins can now use the cost prediction feature and timing optimization option before proceeding with a virtual machine live migration. Testing results show that our proposed approach for VMs live migration cost prediction shows acceptable results with less than 20\% prediction error and can be easily implemented and integrated with VMware vSphere as an example of a commonly used resource management portal for virtual data-centers and private cloud environments. The results show that using our proposed VMs migration timing optimization technique also could save up to 51\% of migration time of the VMs migration time for memory intensive workloads and up to 27\% of the migration time for network intensive workloads. This timing optimization technique can be useful for network admins to save migration time with utilizing higher network rate and higher probability of success. At the end of this thesis, we discuss the persistent memory technology as a new trend in servers memory technology. Persistent memory modes of operation and configurations are discussed in detail to explain how live migration works between servers with different memory configuration set up. Then, we build a VMware cluster with persistent memory inside server and also with DRAM only servers to show the live migration cost difference between the VMs with DRAM only versus the VMs with persistent memory inside.}, language = {en} } @phdthesis{Floeter2005, author = {Fl{\"o}ter, Andr{\´e}}, title = {Analyzing biological expression data based on decision tree induction}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus-6416}, school = {Universit{\"a}t Potsdam}, year = {2005}, abstract = {Modern biological analysis techniques supply scientists with various forms of data. One category of such data are the so called "expression data". These data indicate the quantities of biochemical compounds present in tissue samples. Recently, expression data can be generated at a high speed. This leads in turn to amounts of data no longer analysable by classical statistical techniques. Systems biology is the new field that focuses on the modelling of this information. At present, various methods are used for this purpose. One superordinate class of these meth­ods is machine learning. Methods of this kind had, until recently, predominantly been used for classification and prediction tasks. This neglected a powerful secondary benefit: the ability to induce interpretable models. Obtaining such models from data has become a key issue within Systems biology. Numerous approaches have been proposed and intensively discussed. This thesis focuses on the examination and exploitation of one basic technique: decision trees. The concept of comparing sets of decision trees is developed. This method offers the pos­sibility of identifying significant thresholds in continuous or discrete valued attributes through their corresponding set of decision trees. Finding significant thresholds in attributes is a means of identifying states in living organisms. Knowing about states is an invaluable clue to the un­derstanding of dynamic processes in organisms. Applied to metabolite concentration data, the proposed method was able to identify states which were not found with conventional techniques for threshold extraction. A second approach exploits the structure of sets of decision trees for the discovery of com­binatorial dependencies between attributes. Previous work on this issue has focused either on expensive computational methods or the interpretation of single decision trees ­ a very limited exploitation of the data. This has led to incomplete or unstable results. That is why a new method is developed that uses sets of decision trees to overcome these limitations. Both the introduced methods are available as software tools. They can be applied consecu­tively or separately. That way they make up a package of analytical tools that usefully supplement existing methods. By means of these tools, the newly introduced methods were able to confirm existing knowl­edge and to suggest interesting and new relationships between metabolites.}, subject = {Molekulare Bioinformatik}, language = {en} } @article{FrommholdHeimBarabanovetal.2019, author = {Frommhold, Martin and Heim, Arend and Barabanov, Mikhail and Maier, Franziska and M{\"u}hle, Ralf-Udo and Smirenski, Sergei M. and Heim, Wieland}, title = {Breeding habitat and nest-site selection by an obligatory "nest-cleptoparasite", the Amur Falcon Falco amurensis}, series = {Ecology and evolution}, volume = {9}, journal = {Ecology and evolution}, number = {24}, publisher = {Wiley}, address = {Hoboken}, issn = {2045-7758}, doi = {10.1002/ece3.5878}, pages = {14430 -- 14441}, year = {2019}, abstract = {The selection of a nest site is crucial for successful reproduction of birds. Animals which re-use or occupy nest sites constructed by other species often have limited choice. Little is known about the criteria of nest-stealing species to choose suitable nesting sites and habitats. Here, we analyze breeding-site selection of an obligatory "nest-cleptoparasite", the Amur Falcon Falco amurensis. We collected data on nest sites at Muraviovka Park in the Russian Far East, where the species breeds exclusively in nests of the Eurasian Magpie Pica pica. We sampled 117 Eurasian Magpie nests, 38 of which were occupied by Amur Falcons. Nest-specific variables were assessed, and a recently developed habitat classification map was used to derive landscape metrics. We found that Amur Falcons chose a wide range of nesting sites, but significantly preferred nests with a domed roof. Breeding pairs of Eurasian Hobby Falco subbuteo and Eurasian Magpie were often found to breed near the nest in about the same distance as neighboring Amur Falcon pairs. Additionally, the occurrence of the species was positively associated with bare soil cover, forest cover, and shrub patches within their home range and negatively with the distance to wetlands. Areas of wetlands and fallow land might be used for foraging since Amur Falcons mostly depend on an insect diet. Additionally, we found that rarely burned habitats were preferred. Overall, the effect of landscape variables on the choice of actual nest sites appeared to be rather small. We used different classification methods to predict the probability of occurrence, of which the Random forest method showed the highest accuracy. The areas determined as suitable habitat showed a high concordance with the actual nest locations. We conclude that Amur Falcons prefer to occupy newly built (domed) nests to ensure high nest quality, as well as nests surrounded by available feeding habitats.}, language = {en} } @article{GhafarianWielandLuettschwageretal.2022, author = {Ghafarian, Fatemeh and Wieland, Ralf and L{\"u}ttschwager, Dietmar and Nendel, Claas}, title = {Application of extreme gradient boosting and Shapley Additive explanations to predict temperature regimes inside forests from standard open-field meteorological data}, series = {Environmental modelling \& software with environment data news}, volume = {156}, journal = {Environmental modelling \& software with environment data news}, publisher = {Elsevier}, address = {Oxford}, issn = {1364-8152}, doi = {10.1016/j.envsoft.2022.105466}, pages = {11}, year = {2022}, abstract = {Forest microclimate can buffer biotic responses to summer heat waves, which are expected to become more extreme under climate warming. Prediction of forest microclimate is limited because meteorological observation standards seldom include situations inside forests. We use eXtreme Gradient Boosting - a Machine Learning technique - to predict the microclimate of forest sites in Brandenburg, Germany, using seasonal data comprising weather features. The analysis was amended by applying a SHapley Additive explanation to show the interaction effect of variables and individualised feature attributions. We evaluate model performance in comparison to artificial neural networks, random forest, support vector machine, and multi-linear regression. After implementing a feature selection, an ensemble approach was applied to combine individual models for each forest and improve robustness over a given single prediction model. The resulting model can be applied to translate climate change scenarios into temperatures inside forests to assess temperature-related ecosystem services provided by forests.}, language = {en} } @phdthesis{Haider2013, author = {Haider, Peter}, title = {Prediction with Mixture Models}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus-69617}, school = {Universit{\"a}t Potsdam}, year = {2013}, abstract = {Learning a model for the relationship between the attributes and the annotated labels of data examples serves two purposes. Firstly, it enables the prediction of the label for examples without annotation. Secondly, the parameters of the model can provide useful insights into the structure of the data. If the data has an inherent partitioned structure, it is natural to mirror this structure in the model. Such mixture models predict by combining the individual predictions generated by the mixture components which correspond to the partitions in the data. Often the partitioned structure is latent, and has to be inferred when learning the mixture model. Directly evaluating the accuracy of the inferred partition structure is, in many cases, impossible because the ground truth cannot be obtained for comparison. However it can be assessed indirectly by measuring the prediction accuracy of the mixture model that arises from it. This thesis addresses the interplay between the improvement of predictive accuracy by uncovering latent cluster structure in data, and further addresses the validation of the estimated structure by measuring the accuracy of the resulting predictive model. In the application of filtering unsolicited emails, the emails in the training set are latently clustered into advertisement campaigns. Uncovering this latent structure allows filtering of future emails with very low false positive rates. In order to model the cluster structure, a Bayesian clustering model for dependent binary features is developed in this thesis. Knowing the clustering of emails into campaigns can also aid in uncovering which emails have been sent on behalf of the same network of captured hosts, so-called botnets. This association of emails to networks is another layer of latent clustering. Uncovering this latent structure allows service providers to further increase the accuracy of email filtering and to effectively defend against distributed denial-of-service attacks. To this end, a discriminative clustering model is derived in this thesis that is based on the graph of observed emails. The partitionings inferred using this model are evaluated through their capacity to predict the campaigns of new emails. Furthermore, when classifying the content of emails, statistical information about the sending server can be valuable. Learning a model that is able to make use of it requires training data that includes server statistics. In order to also use training data where the server statistics are missing, a model that is a mixture over potentially all substitutions thereof is developed. Another application is to predict the navigation behavior of the users of a website. Here, there is no a priori partitioning of the users into clusters, but to understand different usage scenarios and design different layouts for them, imposing a partitioning is necessary. The presented approach simultaneously optimizes the discriminative as well as the predictive power of the clusters. Each model is evaluated on real-world data and compared to baseline methods. The results show that explicitly modeling the assumptions about the latent cluster structure leads to improved predictions compared to the baselines. It is beneficial to incorporate a small number of hyperparameters that can be tuned to yield the best predictions in cases where the prediction accuracy can not be optimized directly.}, language = {en} } @article{HampfNendelStreyetal.2021, author = {Hampf, Anna and Nendel, Claas and Strey, Simone and Strey, Robert}, title = {Biotic yield losses in the Southern Amazon, Brazil}, series = {Frontiers in plant science : FPLS}, volume = {12}, journal = {Frontiers in plant science : FPLS}, publisher = {Frontiers Media}, address = {Lausanne}, issn = {1664-462X}, doi = {10.3389/fpls.2021.621168}, pages = {16}, year = {2021}, abstract = {Pathogens and animal pests (P\&A) are a major threat to global food security as they directly affect the quantity and quality of food. The Southern Amazon, Brazil's largest domestic region for soybean, maize and cotton production, is particularly vulnerable to the outbreak of P\&A due to its (sub)tropical climate and intensive farming systems. However, little is known about the spatial distribution of P\&A and the related yield losses. Machine learning approaches for the automated recognition of plant diseases can help to overcome this research gap. The main objectives of this study are to (1) evaluate the performance of Convolutional Neural Networks (ConvNets) in classifying P\&A, (2) map the spatial distribution of P\&A in the Southern Amazon, and (3) quantify perceived yield and economic losses for the main soybean and maize P\&A. The objectives were addressed by making use of data collected with the smartphone application Plantix. The core of the app's functioning is the automated recognition of plant diseases via ConvNets. Data on expected yield losses were gathered through a short survey included in an "expert" version of the application, which was distributed among agronomists. Between 2016 and 2020, Plantix users collected approximately 78,000 georeferenced P\&A images in the Southern Amazon. The study results indicate a high performance of the trained ConvNets in classifying 420 different crop-disease combinations. Spatial distribution maps and expert-based yield loss estimates indicate that maize rust, bacterial stalk rot and the fall armyworm are among the most severe maize P\&A, whereas soybean is mainly affected by P\&A like anthracnose, downy mildew, frogeye leaf spot, stink bugs and brown spot. Perceived soybean and maize yield losses amount to 12 and 16\%, respectively, resulting in annual yield losses of approximately 3.75 million tonnes for each crop and economic losses of US\$2 billion for both crops together. The high level of accuracy of the trained ConvNets, when paired with widespread use from following a citizen-science approach, results in a data source that will shed new light on yield loss estimates, e.g., for the analysis of yield gaps and the development of measures to minimise them.}, language = {en} } @article{HeckerSteckhanEybenetal.2022, author = {Hecker, Pascal and Steckhan, Nico and Eyben, Florian and Schuller, Bj{\"o}rn Wolfgang and Arnrich, Bert}, title = {Voice Analysis for Neurological Disorder Recognition - A Systematic Review and Perspective on Emerging Trends}, series = {Frontiers in Digital Health}, journal = {Frontiers in Digital Health}, publisher = {Frontiers Media SA}, address = {Lausanne, Schweiz}, issn = {2673-253X}, doi = {10.3389/fdgth.2022.842301}, pages = {16}, year = {2022}, abstract = {Quantifying neurological disorders from voice is a rapidly growing field of research and holds promise for unobtrusive and large-scale disorder monitoring. The data recording setup and data analysis pipelines are both crucial aspects to effectively obtain relevant information from participants. Therefore, we performed a systematic review to provide a high-level overview of practices across various neurological disorders and highlight emerging trends. PRISMA-based literature searches were conducted through PubMed, Web of Science, and IEEE Xplore to identify publications in which original (i.e., newly recorded) datasets were collected. Disorders of interest were psychiatric as well as neurodegenerative disorders, such as bipolar disorder, depression, and stress, as well as amyotrophic lateral sclerosis amyotrophic lateral sclerosis, Alzheimer's, and Parkinson's disease, and speech impairments (aphasia, dysarthria, and dysphonia). Of the 43 retrieved studies, Parkinson's disease is represented most prominently with 19 discovered datasets. Free speech and read speech tasks are most commonly used across disorders. Besides popular feature extraction toolkits, many studies utilise custom-built feature sets. Correlations of acoustic features with psychiatric and neurodegenerative disorders are presented. In terms of analysis, statistical analysis for significance of individual features is commonly used, as well as predictive modeling approaches, especially with support vector machines and a small number of artificial neural networks. An emerging trend and recommendation for future studies is to collect data in everyday life to facilitate longitudinal data collection and to capture the behavior of participants more naturally. Another emerging trend is to record additional modalities to voice, which can potentially increase analytical performance.}, language = {en} } @misc{HeckerSteckhanEybenetal.2022, author = {Hecker, Pascal and Steckhan, Nico and Eyben, Florian and Schuller, Bj{\"o}rn Wolfgang and Arnrich, Bert}, title = {Voice Analysis for Neurological Disorder Recognition - A Systematic Review and Perspective on Emerging Trends}, series = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Reihe der Digital Engineering Fakult{\"a}t}, journal = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Reihe der Digital Engineering Fakult{\"a}t}, number = {13}, doi = {10.25932/publishup-58101}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-581019}, pages = {16}, year = {2022}, abstract = {Quantifying neurological disorders from voice is a rapidly growing field of research and holds promise for unobtrusive and large-scale disorder monitoring. The data recording setup and data analysis pipelines are both crucial aspects to effectively obtain relevant information from participants. Therefore, we performed a systematic review to provide a high-level overview of practices across various neurological disorders and highlight emerging trends. PRISMA-based literature searches were conducted through PubMed, Web of Science, and IEEE Xplore to identify publications in which original (i.e., newly recorded) datasets were collected. Disorders of interest were psychiatric as well as neurodegenerative disorders, such as bipolar disorder, depression, and stress, as well as amyotrophic lateral sclerosis amyotrophic lateral sclerosis, Alzheimer's, and Parkinson's disease, and speech impairments (aphasia, dysarthria, and dysphonia). Of the 43 retrieved studies, Parkinson's disease is represented most prominently with 19 discovered datasets. Free speech and read speech tasks are most commonly used across disorders. Besides popular feature extraction toolkits, many studies utilise custom-built feature sets. Correlations of acoustic features with psychiatric and neurodegenerative disorders are presented. In terms of analysis, statistical analysis for significance of individual features is commonly used, as well as predictive modeling approaches, especially with support vector machines and a small number of artificial neural networks. An emerging trend and recommendation for future studies is to collect data in everyday life to facilitate longitudinal data collection and to capture the behavior of participants more naturally. Another emerging trend is to record additional modalities to voice, which can potentially increase analytical performance.}, language = {en} } @phdthesis{Hoang2019, author = {Hoang, Yen}, title = {De novo binning strategy to analyze and visualize multi-dimensional cytometric data}, doi = {10.25932/publishup-44307}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-443078}, school = {Universit{\"a}t Potsdam}, pages = {vii, 81, xxxii}, year = {2019}, abstract = {Since half a century, cytometry has been a major scientific discipline in the field of cytomics - the study of system's biology at single cell level. It enables the investigation of physiological processes, functional characteristics and rare events with proteins by analysing multiple parameters on an individual cell basis. In the last decade, mass cytometry has been established which increased the parallel measurement to up to 50 proteins. This has shifted the analysis strategy from conventional consecutive manual gates towards multi-dimensional data processing. Novel algorithms have been developed to tackle these high-dimensional protein combinations in the data. They are mainly based on clustering or non-linear dimension reduction techniques, or both, often combined with an upstream downsampling procedure. However, these tools have obstacles either in comprehensible interpretability, reproducibility, computational complexity or in comparability between samples and groups. To address this bottleneck, a reproducible, semi-automated cytometric data mining workflow PRI (pattern recognition of immune cells) is proposed which combines three main steps: i) data preparation and storage; ii) bin-based combinatorial variable engineering of three protein markers, the so called triploTs, and subsequent sectioning of these triploTs in four parts; and iii) deployment of a data-driven supervised learning algorithm, the cross-validated elastic-net regularized logistic regression, with these triploT sections as input variables. As a result, the selected variables from the models are ranked by their prevalence, which potentially have discriminative value. The purpose is to significantly facilitate the identification of meaningful subpopulations, which are most distinguish between two groups. The proposed workflow PRI is exemplified by a recently published public mass cytometry data set. The authors found a T cell subpopulation which is discriminative between effective and ineffective treatment of breast carcinomas in mice. With PRI, that subpopulation was not only validated, but was further narrowed down as a particular Th1 cell population. Moreover, additional insights of combinatorial protein expressions are revealed in a traceable manner. An essential element in the workflow is the reproducible variable engineering. These variables serve as basis for a clearly interpretable visualization, for a structured variable exploration and as input layers in neural network constructs. PRI facilitates the determination of marker levels in a semi-continuous manner. Jointly with the combinatorial display, it allows a straightforward observation of correlating patterns, and thus, the dominant expressed markers and cell hierarchies. Furthermore, it enables the identification and complex characterization of discriminating subpopulations due to its reproducible and pseudo-multi-parametric pattern presentation. This endorses its applicability as a tool for unbiased investigations on cell subsets within multi-dimensional cytometric data sets.}, language = {en} } @misc{HollsteinSeglGuanteretal.2016, author = {Hollstein, Andr{\´e} and Segl, Karl and Guanter, Luis and Brell, Maximilian and Enesco, Marta}, title = {Ready-to-Use methods for the detection of clouds, cirrus, snow, shadow, water and clear sky pixels in Sentinel-2 MSI images}, series = {remote sensing}, journal = {remote sensing}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-407938}, pages = {18}, year = {2016}, abstract = {Classification of clouds, cirrus, snow, shadows and clear sky areas is a crucial step in the pre-processing of optical remote sensing images and is a valuable input for their atmospheric correction. The Multi-Spectral Imager on board the Sentinel-2's of the Copernicus program offers optimized bands for this task and delivers unprecedented amounts of data regarding spatial sampling, global coverage, spectral coverage, and repetition rate. Efficient algorithms are needed to process, or possibly reprocess, those big amounts of data. Techniques based on top-of-atmosphere reflectance spectra for single-pixels without exploitation of external data or spatial context offer the largest potential for parallel data processing and highly optimized processing throughput. Such algorithms can be seen as a baseline for possible trade-offs in processing performance when the application of more sophisticated methods is discussed. We present several ready-to-use classification algorithms which are all based on a publicly available database of manually classified Sentinel-2A images. These algorithms are based on commonly used and newly developed machine learning techniques which drastically reduce the amount of time needed to update the algorithms when new images are added to the database. Several ready-to-use decision trees are presented which allow to correctly label about 91\% of the spectra within a validation dataset. While decision trees are simple to implement and easy to understand, they offer only limited classification skill. It improves to 98\% when the presented algorithm based on the classical Bayesian method is applied. This method has only recently been used for this task and shows excellent performance concerning classification skill and processing performance. A comparison of the presented algorithms with other commonly used techniques such as random forests, stochastic gradient descent, or support vector machines is also given. Especially random forests and support vector machines show similar classification skill as the classical Bayesian method.}, language = {en} } @article{KibrikKhudyakovaDobrovetal.2016, author = {Kibrik, Andrej A. and Khudyakova, Mariya V. and Dobrov, Grigory B. and Linnik, Anastasia and Zalmanov, Dmitrij A.}, title = {Referential Choice}, series = {Frontiers in psychology}, volume = {7}, journal = {Frontiers in psychology}, publisher = {Frontiers Research Foundation}, address = {Lausanne}, issn = {1664-1078}, doi = {10.3389/fpsyg.2016.01429}, year = {2016}, abstract = {We report a study of referential choice in discourse production, understood as the choice between various types of referential devices, such as pronouns and full noun phrases. Our goal is to predict referential choice, and to explore to what extent such prediction is possible. Our approach to referential choice includes a cognitively informed theoretical component, corpus analysis, machine learning methods and experimentation with human participants. Machine learning algorithms make use of 25 factors, including referent's properties (such as animacy and protagonism), the distance between a referential expression and its antecedent, the antecedent's syntactic role, and so on. Having found the predictions of our algorithm to coincide with the original almost 90\% of the time, we hypothesized that fully accurate prediction is not possible because, in many situations, more than one referential option is available. This hypothesis was supported by an experimental study, in which participants answered questions about either the original text in the corpus, or about a text modified in accordance with the algorithm's prediction. Proportions of correct answers to these questions, as well as participants' rating of the questions' difficulty, suggested that divergences between the algorithm's prediction and the original referential device in the corpus occur overwhelmingly in situations where the referential choice is not categorical.}, language = {en} } @misc{KibrikKhudyakovaDobrovetal.2016, author = {Kibrik, Andrej A. and Khudyakova, Mariya V. and Dobrov, Grigory B. and Linnik, Anastasia and Zalmanov, Dmitrij A.}, title = {Referential Choice}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-100313}, pages = {21}, year = {2016}, abstract = {We report a study of referential choice in discourse production, understood as the choice between various types of referential devices, such as pronouns and full noun phrases. Our goal is to predict referential choice, and to explore to what extent such prediction is possible. Our approach to referential choice includes a cognitively informed theoretical component, corpus analysis, machine learning methods and experimentation with human participants. Machine learning algorithms make use of 25 factors, including referent's properties (such as animacy and protagonism), the distance between a referential expression and its antecedent, the antecedent's syntactic role, and so on. Having found the predictions of our algorithm to coincide with the original almost 90\% of the time, we hypothesized that fully accurate prediction is not possible because, in many situations, more than one referential option is available. This hypothesis was supported by an experimental study, in which participants answered questions about either the original text in the corpus, or about a text modified in accordance with the algorithm's prediction. Proportions of correct answers to these questions, as well as participants' rating of the questions' difficulty, suggested that divergences between the algorithm's prediction and the original referential device in the corpus occur overwhelmingly in situations where the referential choice is not categorical.}, language = {en} } @article{KibrikKhudyakovaDobrovetal.2016, author = {Kibrik, Andrej A. and Khudyakova, Mariya V. and Dobrov, Grigory B. and Linnik, Anastasia and Zalmanov, Dmitrij A.}, title = {Referential Choice: Predictability and Its Limits}, series = {Frontiers in psychology}, volume = {7}, journal = {Frontiers in psychology}, publisher = {Frontiers Research Foundation}, address = {Lausanne}, issn = {1664-1078}, doi = {10.3389/fpsyg.2016.01429}, pages = {9939 -- 9947}, year = {2016}, abstract = {We report a study of referential choice in discourse production, understood as the choice between various types of referential devices, such as pronouns and full noun phrases. Our goal is to predict referential choice, and to explore to what extent such prediction is possible. Our approach to referential choice includes a cognitively informed theoretical component, corpus analysis, machine learning methods and experimentation with human participants. Machine learning algorithms make use of 25 factors, including referent's properties (such as animacy and protagonism), the distance between a referential expression and its antecedent, the antecedent's syntactic role, and so on. Having found the predictions of our algorithm to coincide with the original almost 90\% of the time, we hypothesized that fully accurate prediction is not possible because, in many situations, more than one referential option is available. This hypothesis was supported by an experimental study, in which participants answered questions about either the original text in the corpus, or about a text modified in accordance with the algorithm's prediction. Proportions of correct answers to these questions, as well as participants' rating of the questions' difficulty, suggested that divergences between the algorithm's prediction and the original referential device in the corpus occur overwhelmingly in situations where the referential choice is not categorical.}, language = {en} } @article{KonakWegnerArnrich2020, author = {Konak, Orhan and Wegner, Pit and Arnrich, Bert}, title = {IMU-Based Movement Trajectory Heatmaps for Human Activity Recognition}, series = {Sensors}, volume = {20}, journal = {Sensors}, number = {24}, publisher = {MDPI}, address = {Basel}, issn = {1424-8220}, doi = {10.3390/s20247179}, pages = {15}, year = {2020}, abstract = {Recent trends in ubiquitous computing have led to a proliferation of studies that focus on human activity recognition (HAR) utilizing inertial sensor data that consist of acceleration, orientation and angular velocity. However, the performances of such approaches are limited by the amount of annotated training data, especially in fields where annotating data is highly time-consuming and requires specialized professionals, such as in healthcare. In image classification, this limitation has been mitigated by powerful oversampling techniques such as data augmentation. Using this technique, this work evaluates to what extent transforming inertial sensor data into movement trajectories and into 2D heatmap images can be advantageous for HAR when data are scarce. A convolutional long short-term memory (ConvLSTM) network that incorporates spatiotemporal correlations was used to classify the heatmap images. Evaluation was carried out on Deep Inertial Poser (DIP), a known dataset composed of inertial sensor data. The results obtained suggest that for datasets with large numbers of subjects, using state-of-the-art methods remains the best alternative. However, a performance advantage was achieved for small datasets, which is usually the case in healthcare. Moreover, movement trajectories provide a visual representation of human activities, which can help researchers to better interpret and analyze motion patterns.}, language = {en} } @misc{KonakWegnerArnrich2021, author = {Konak, Orhan and Wegner, Pit and Arnrich, Bert}, title = {IMU-Based Movement Trajectory Heatmaps for Human Activity Recognition}, series = {Postprints der Universit{\"a}t Potsdam : Reihe der Digital Engineering Fakult{\"a}t}, journal = {Postprints der Universit{\"a}t Potsdam : Reihe der Digital Engineering Fakult{\"a}t}, number = {4}, doi = {10.25932/publishup-48779}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-487799}, pages = {17}, year = {2021}, abstract = {Recent trends in ubiquitous computing have led to a proliferation of studies that focus on human activity recognition (HAR) utilizing inertial sensor data that consist of acceleration, orientation and angular velocity. However, the performances of such approaches are limited by the amount of annotated training data, especially in fields where annotating data is highly time-consuming and requires specialized professionals, such as in healthcare. In image classification, this limitation has been mitigated by powerful oversampling techniques such as data augmentation. Using this technique, this work evaluates to what extent transforming inertial sensor data into movement trajectories and into 2D heatmap images can be advantageous for HAR when data are scarce. A convolutional long short-term memory (ConvLSTM) network that incorporates spatiotemporal correlations was used to classify the heatmap images. Evaluation was carried out on Deep Inertial Poser (DIP), a known dataset composed of inertial sensor data. The results obtained suggest that for datasets with large numbers of subjects, using state-of-the-art methods remains the best alternative. However, a performance advantage was achieved for small datasets, which is usually the case in healthcare. Moreover, movement trajectories provide a visual representation of human activities, which can help researchers to better interpret and analyze motion patterns.}, language = {en} } @phdthesis{Kotha2018, author = {Kotha, Sreeram Reddy}, title = {Quantification of uncertainties in seismic ground-motion prediction}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-415743}, school = {Universit{\"a}t Potsdam}, pages = {xii, 101}, year = {2018}, abstract = {The purpose of Probabilistic Seismic Hazard Assessment (PSHA) at a construction site is to provide the engineers with a probabilistic estimate of ground-motion level that could be equaled or exceeded at least once in the structure's design lifetime. A certainty on the predicted ground-motion allows the engineers to confidently optimize structural design and mitigate the risk of extensive damage, or in worst case, a collapse. It is therefore in interest of engineering, insurance, disaster mitigation, and security of society at large, to reduce uncertainties in prediction of design ground-motion levels. In this study, I am concerned with quantifying and reducing the prediction uncertainty of regression-based Ground-Motion Prediction Equations (GMPEs). Essentially, GMPEs are regressed best-fit formulae relating event, path, and site parameters (predictor variables) to observed ground-motion values at the site (prediction variable). GMPEs are characterized by a parametric median (μ) and a non-parametric variance (σ) of prediction. μ captures the known ground-motion physics i.e., scaling with earthquake rupture properties (event), attenuation with distance from source (region/path), and amplification due to local soil conditions (site); while σ quantifies the natural variability of data that eludes μ. In a broad sense, the GMPE prediction uncertainty is cumulative of 1) uncertainty on estimated regression coefficients (uncertainty on μ,σ_μ), and 2) the inherent natural randomness of data (σ). The extent of μ parametrization, the quantity, and quality of ground-motion data used in a regression, govern the size of its prediction uncertainty: σ_μ and σ. In the first step, I present the impact of μ parametrization on the size of σ_μ and σ. Over-parametrization appears to increase the σ_μ, because of the large number of regression coefficients (in μ) to be estimated with insufficient data. Under-parametrization mitigates σ_μ, but the reduced explanatory strength of μ is reflected in inflated σ. For an optimally parametrized GMPE, a ~10\% reduction in σ is attained by discarding the low-quality data from pan-European events with incorrect parametric values (of predictor variables). In case of regions with scarce ground-motion recordings, without under-parametrization, the only way to mitigate σ_μ is to substitute long-term earthquake data at a location with short-term samples of data across several locations - the Ergodic Assumption. However, the price of ergodic assumption is an increased σ, due to the region-to-region and site-to-site differences in ground-motion physics. σ of an ergodic GMPE developed from generic ergodic dataset is much larger than that of non-ergodic GMPEs developed from region- and site-specific non-ergodic subsets - which were too sparse to produce their specific GMPEs. Fortunately, with the dramatic increase in recorded ground-motion data at several sites across Europe and Middle-East, I could quantify the region- and site-specific differences in ground-motion scaling and upgrade the GMPEs with 1) substantially more accurate region- and site-specific μ for sites in Italy and Turkey, and 2) significantly smaller prediction variance σ. The benefit of such enhancements to GMPEs is quite evident in my comparison of PSHA estimates from ergodic versus region- and site-specific GMPEs; where the differences in predicted design ground-motion levels, at several sites in Europe and Middle-Eastern regions, are as large as ~50\%. Resolving the ergodic assumption with mixed-effects regressions is feasible when the quantified region- and site-specific effects are physically meaningful, and the non-ergodic subsets (regions and sites) are defined a priori through expert knowledge. In absence of expert definitions, I demonstrate the potential of machine learning techniques in identifying efficient clusters of site-specific non-ergodic subsets, based on latent similarities in their ground-motion data. Clustered site-specific GMPEs bridge the gap between site-specific and fully ergodic GMPEs, with their partially non-ergodic μ and, σ ~15\% smaller than the ergodic variance. The methodological refinements to GMPE development produced in this study are applicable to new ground-motion datasets, to further enhance certainty of ground-motion prediction and thereby, seismic hazard assessment. Advanced statistical tools show great potential in improving the predictive capabilities of GMPEs, but the fundamental requirement remains: large quantity of high-quality ground-motion data from several sites for an extended time-period.}, language = {en} } @phdthesis{Koumarelas2020, author = {Koumarelas, Ioannis}, title = {Data preparation and domain-agnostic duplicate detection}, doi = {10.25932/publishup-48913}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-489131}, school = {Universit{\"a}t Potsdam}, pages = {x, 97}, year = {2020}, abstract = {Successfully completing any data science project demands careful consideration across its whole process. Although the focus is often put on later phases of the process, in practice, experts spend more time in earlier phases, preparing data, to make them consistent with the systems' requirements or to improve their models' accuracies. Duplicate detection is typically applied during the data cleaning phase, which is dedicated to removing data inconsistencies and improving the overall quality and usability of data. While data cleaning involves a plethora of approaches to perform specific operations, such as schema alignment and data normalization, the task of detecting and removing duplicate records is particularly challenging. Duplicates arise when multiple records representing the same entities exist in a database. Due to numerous reasons, spanning from simple typographical errors to different schemas and formats of integrated databases. Keeping a database free of duplicates is crucial for most use-cases, as their existence causes false negatives and false positives when matching queries against it. These two data quality issues have negative implications for tasks, such as hotel booking, where users may erroneously select a wrong hotel, or parcel delivery, where a parcel can get delivered to the wrong address. Identifying the variety of possible data issues to eliminate duplicates demands sophisticated approaches. While research in duplicate detection is well-established and covers different aspects of both efficiency and effectiveness, our work in this thesis focuses on the latter. We propose novel approaches to improve data quality before duplicate detection takes place and apply the latter in datasets even when prior labeling is not available. Our experiments show that improving data quality upfront can increase duplicate classification results by up to 19\%. To this end, we propose two novel pipelines that select and apply generic as well as address-specific data preparation steps with the purpose of maximizing the success of duplicate detection. Generic data preparation, such as the removal of special characters, can be applied to any relation with alphanumeric attributes. When applied, data preparation steps are selected only for attributes where there are positive effects on pair similarities, which indirectly affect classification, or on classification directly. Our work on addresses is twofold; first, we consider more domain-specific approaches to improve the quality of values, and, second, we experiment with known and modified versions of similarity measures to select the most appropriate per address attribute, e.g., city or country. To facilitate duplicate detection in applications where gold standard annotations are not available and obtaining them is not possible or too expensive, we propose MDedup. MDedup is a novel, rule-based, and fully automatic duplicate detection approach that is based on matching dependencies. These dependencies can be used to detect duplicates and can be discovered using state-of-the-art algorithms efficiently and without any prior labeling. MDedup uses two pipelines to first train on datasets with known labels, learning to identify useful matching dependencies, and then be applied on unseen datasets, regardless of any existing gold standard. Finally, our work is accompanied by open source code to enable repeatability of our research results and application of our approaches to other datasets.}, language = {en} } @article{KuehnHainzlDahmetal.2022, author = {K{\"u}hn, Daniela and Hainzl, Sebastian and Dahm, Torsten and Richter, Gudrun and Vera Rodriguez, Ismael}, title = {A review of source models to further the understanding of the seismicity of the Groningen field}, series = {Netherlands journal of geosciences : NJG}, volume = {101}, journal = {Netherlands journal of geosciences : NJG}, publisher = {Cambridge Univ. Press}, address = {Cambridge}, issn = {0016-7746}, doi = {10.1017/njg.2022.7}, pages = {12}, year = {2022}, abstract = {The occurrence of felt earthquakes due to gas production in Groningen has initiated numerous studies and model attempts to understand and quantify induced seismicity in this region. The whole bandwidth of available models spans the range from fully deterministic models to purely empirical and stochastic models. In this article, we summarise the most important model approaches, describing their main achievements and limitations. In addition, we discuss remaining open questions and potential future directions of development.}, language = {en} } @phdthesis{Lazaridou2021, author = {Lazaridou, Konstantina}, title = {Revealing hidden patterns in political news and social media with machine learning}, doi = {10.25932/publishup-50273}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-502734}, school = {Universit{\"a}t Potsdam}, pages = {xv, 140}, year = {2021}, abstract = {As part of our everyday life we consume breaking news and interpret it based on our own viewpoints and beliefs. We have easy access to online social networking platforms and news media websites, where we inform ourselves about current affairs and often post about our own views, such as in news comments or social media posts. The media ecosystem enables opinions and facts to travel from news sources to news readers, from news article commenters to other readers, from social network users to their followers, etc. The views of the world many of us have depend on the information we receive via online news and social media. Hence, it is essential to maintain accurate, reliable and objective online content to ensure democracy and verity on the Web. To this end, we contribute to a trustworthy media ecosystem by analyzing news and social media in the context of politics to ensure that media serves the public interest. In this thesis, we use text mining, natural language processing and machine learning techniques to reveal underlying patterns in political news articles and political discourse in social networks. Mainstream news sources typically cover a great amount of the same news stories every day, but they often place them in a different context or report them from different perspectives. In this thesis, we are interested in how distinct and predictable newspaper journalists are, in the way they report the news, as a means to understand and identify their different political beliefs. To this end, we propose two models that classify text from news articles to their respective original news source, i.e., reported speech and also news comments. Our goal is to capture systematic quoting and commenting patterns by journalists and news commenters respectively, which can lead us to the newspaper where the quotes and comments are originally published. Predicting news sources can help us understand the potential subjective nature behind news storytelling and the magnitude of this phenomenon. Revealing this hidden knowledge can restore our trust in media by advancing transparency and diversity in the news. Media bias can be expressed in various subtle ways in the text and it is often challenging to identify these bias manifestations correctly, even for humans. However, media experts, e.g., journalists, are a powerful resource that can help us overcome the vague definition of political media bias and they can also assist automatic learners to find the hidden bias in the text. Due to the enormous technological advances in artificial intelligence, we hypothesize that identifying political bias in the news could be achieved through the combination of sophisticated deep learning modelsxi and domain expertise. Therefore, our second contribution is a high-quality and reliable news dataset annotated by journalists for political bias and a state-of-the-art solution for this task based on curriculum learning. Our aim is to discover whether domain expertise is necessary for this task and to provide an automatic solution for this traditionally manually-solved problem. User generated content is fundamentally different from news articles, e.g., messages are shorter, they are often personal and opinionated, they refer to specific topics and persons, etc. Regarding political and socio-economic news, individuals in online communities make use of social networks to keep their peers up-to-date and to share their own views on ongoing affairs. We believe that social media is also an as powerful instrument for information flow as the news sources are, and we use its unique characteristic of rapid news coverage for two applications. We analyze Twitter messages and debate transcripts during live political presidential debates to automatically predict the topics that Twitter users discuss. Our goal is to discover the favoured topics in online communities on the dates of political events as a way to understand the political subjects of public interest. With the up-to-dateness of microblogs, an additional opportunity emerges, namely to use social media posts and leverage the real-time verity about discussed individuals to find their locations. That is, given a person of interest that is mentioned in online discussions, we use the wisdom of the crowd to automatically track her physical locations over time. We evaluate our approach in the context of politics, i.e., we predict the locations of US politicians as a proof of concept for important use cases, such as to track people that are national risks, e.g., warlords and wanted criminals.}, language = {en} } @article{LevyMussackBrunneretal.2020, author = {Levy, Jessica and Mussack, Dominic and Brunner, Martin and Keller, Ulrich and Cardoso-Leite, Pedro and Fischbach, Antoine}, title = {Contrasting classical and machine learning approaches in the estimation of value-added scores in large-scale educational data}, series = {Frontiers in psychology}, volume = {11}, journal = {Frontiers in psychology}, publisher = {Frontiers Research Foundation}, address = {Lausanne}, issn = {1664-1078}, doi = {10.3389/fpsyg.2020.02190}, pages = {18}, year = {2020}, abstract = {There is no consensus on which statistical model estimates school value-added (VA) most accurately. To date, the two most common statistical models used for the calculation of VA scores are two classical methods: linear regression and multilevel models. These models have the advantage of being relatively transparent and thus understandable for most researchers and practitioners. However, these statistical models are bound to certain assumptions (e.g., linearity) that might limit their prediction accuracy. Machine learning methods, which have yielded spectacular results in numerous fields, may be a valuable alternative to these classical models. Although big data is not new in general, it is relatively new in the realm of social sciences and education. New types of data require new data analytical approaches. Such techniques have already evolved in fields with a long tradition in crunching big data (e.g., gene technology). The objective of the present paper is to competently apply these "imported" techniques to education data, more precisely VA scores, and assess when and how they can extend or replace the classical psychometrics toolbox. The different models include linear and non-linear methods and extend classical models with the most commonly used machine learning methods (i.e., random forest, neural networks, support vector machines, and boosting). We used representative data of 3,026 students in 153 schools who took part in the standardized achievement tests of the Luxembourg School Monitoring Program in grades 1 and 3. Multilevel models outperformed classical linear and polynomial regressions, as well as different machine learning models. However, it could be observed that across all schools, school VA scores from different model types correlated highly. Yet, the percentage of disagreements as compared to multilevel models was not trivial and real-life implications for individual schools may still be dramatic depending on the model type used. Implications of these results and possible ethical concerns regarding the use of machine learning methods for decision-making in education are discussed.}, language = {en} } @phdthesis{Lilienkamp2024, author = {Lilienkamp, Henning}, title = {Enhanced computational approaches for data-driven characterization of earthquake ground motion and rapid earthquake impact assessment}, doi = {10.25932/publishup-63195}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-631954}, school = {Universit{\"a}t Potsdam}, pages = {x, 145}, year = {2024}, abstract = {Rapidly growing seismic and macroseismic databases and simplified access to advanced machine learning methods have in recent years opened up vast opportunities to address challenges in engineering and strong motion seismology from novel, datacentric perspectives. In this thesis, I explore the opportunities of such perspectives for the tasks of ground motion modeling and rapid earthquake impact assessment, tasks with major implications for long-term earthquake disaster mitigation. In my first study, I utilize the rich strong motion database from the Kanto basin, Japan, and apply the U-Net artificial neural network architecture to develop a deep learning based ground motion model. The operational prototype provides statistical estimates of expected ground shaking, given descriptions of a specific earthquake source, wave propagation paths, and geophysical site conditions. The U-Net interprets ground motion data in its spatial context, potentially taking into account, for example, the geological properties in the vicinity of observation sites. Predictions of ground motion intensity are thereby calibrated to individual observation sites and earthquake locations. The second study addresses the explicit incorporation of rupture forward directivity into ground motion modeling. Incorporation of this phenomenon, causing strong, pulse like ground shaking in the vicinity of earthquake sources, is usually associated with an intolerable increase in computational demand during probabilistic seismic hazard analysis (PSHA) calculations. I suggest an approach in which I utilize an artificial neural network to efficiently approximate the average, directivity-related adjustment to ground motion predictions for earthquake ruptures from the 2022 New Zealand National Seismic Hazard Model. The practical implementation in an actual PSHA calculation demonstrates the efficiency and operational readiness of my model. In a follow-up study, I present a proof of concept for an alternative strategy in which I target the generalizing applicability to ruptures other than those from the New Zealand National Seismic Hazard Model. In the third study, I address the usability of pseudo-intensity reports obtained from macroseismic observations by non-expert citizens for rapid impact assessment. I demonstrate that the statistical properties of pseudo-intensity collections describing the intensity of shaking are correlated with the societal impact of earthquakes. In a second step, I develop a probabilistic model that, within minutes of an event, quantifies the probability of an earthquake to cause considerable societal impact. Under certain conditions, such a quick and preliminary method might be useful to support decision makers in their efforts to organize auxiliary measures for earthquake disaster response while results from more elaborate impact assessment frameworks are not yet available. The application of machine learning methods to datasets that only partially reveal characteristics of Big Data, qualify the majority of results obtained in this thesis as explorative insights rather than ready-to-use solutions to real world problems. The practical usefulness of this work will be better assessed in the future by applying the approaches developed to growing and increasingly complex data sets.}, language = {en} } @phdthesis{Loster2021, author = {Loster, Michael}, title = {Knowledge base construction with machine learning methods}, doi = {10.25932/publishup-50145}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-501459}, school = {Universit{\"a}t Potsdam}, pages = {ii, 130}, year = {2021}, abstract = {Modern knowledge bases contain and organize knowledge from many different topic areas. Apart from specific entity information, they also store information about their relationships amongst each other. Combining this information results in a knowledge graph that can be particularly helpful in cases where relationships are of central importance. Among other applications, modern risk assessment in the financial sector can benefit from the inherent network structure of such knowledge graphs by assessing the consequences and risks of certain events, such as corporate insolvencies or fraudulent behavior, based on the underlying network structure. As public knowledge bases often do not contain the necessary information for the analysis of such scenarios, the need arises to create and maintain dedicated domain-specific knowledge bases. This thesis investigates the process of creating domain-specific knowledge bases from structured and unstructured data sources. In particular, it addresses the topics of named entity recognition (NER), duplicate detection, and knowledge validation, which represent essential steps in the construction of knowledge bases. As such, we present a novel method for duplicate detection based on a Siamese neural network that is able to learn a dataset-specific similarity measure which is used to identify duplicates. Using the specialized network architecture, we design and implement a knowledge transfer between two deduplication networks, which leads to significant performance improvements and a reduction of required training data. Furthermore, we propose a named entity recognition approach that is able to identify company names by integrating external knowledge in the form of dictionaries into the training process of a conditional random field classifier. In this context, we study the effects of different dictionaries on the performance of the NER classifier. We show that both the inclusion of domain knowledge as well as the generation and use of alias names results in significant performance improvements. For the validation of knowledge represented in a knowledge base, we introduce Colt, a framework for knowledge validation based on the interactive quality assessment of logical rules. In its most expressive implementation, we combine Gaussian processes with neural networks to create Colt-GP, an interactive algorithm for learning rule models. Unlike other approaches, Colt-GP uses knowledge graph embeddings and user feedback to cope with data quality issues of knowledge bases. The learned rule model can be used to conditionally apply a rule and assess its quality. Finally, we present CurEx, a prototypical system for building domain-specific knowledge bases from structured and unstructured data sources. Its modular design is based on scalable technologies, which, in addition to processing large datasets, ensures that the modules can be easily exchanged or extended. CurEx offers multiple user interfaces, each tailored to the individual needs of a specific user group and is fully compatible with the Colt framework, which can be used as part of the system. We conduct a wide range of experiments with different datasets to determine the strengths and weaknesses of the proposed methods. To ensure the validity of our results, we compare the proposed methods with competing approaches.}, language = {en} } @phdthesis{Meier2017, author = {Meier, Sebastian}, title = {Personal Big Data}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-406696}, school = {Universit{\"a}t Potsdam}, pages = {xxiv, 133}, year = {2017}, abstract = {Many users of cloud-based services are concerned about questions of data privacy. At the same time, they want to benefit from smart data-driven services, which require insight into a person's individual behaviour. The modus operandi of user modelling is that data is sent to a remote server where the model is constructed and merged with other users' data. This thesis proposes selective cloud computing, an alternative approach, in which the user model is constructed on the client-side and only an abstracted generalised version of the model is shared with the remote services. In order to demonstrate the applicability of this approach, the thesis builds an exemplary client-side user modelling technique. As this thesis is carried out in the area of Geoinformatics and spatio-temporal data is particularly sensitive, the application domain for this experiment is the analysis and prediction of a user's spatio-temporal behaviour. The user modelling technique is grounded in an innovative conceptual model, which builds upon spatial network theory combined with time-geography. The spatio-temporal constraints of time-geography are applied to the network structure in order to create individual spatio-temporal action spaces. This concept is translated into a novel algorithmic user modelling approach which is solely driven by the user's own spatio-temporal trajectory data that is generated by the user's smartphone. While modern smartphones offer a rich variety of sensory data, this thesis only makes use of spatio-temporal trajectory data, enriched by activity classification, as the input and foundation for the algorithmic model. The algorithmic model consists of three basal components: locations (vertices), trips (edges), and clusters (neighbourhoods). After preprocessing the incoming trajectory data in order to identify locations, user feedback is used to train an artificial neural network to learn temporal patterns for certain location types (e.g. work, home, bus stop, etc.). This Artificial Neural Network (ANN) is used to automatically detect future location types by their spatio-temporal patterns. The same is done in order to predict the duration of stay at a certain location. Experiments revealed that neural nets were the most successful statistical and machine learning tool to detect those patterns. The location type identification algorithm reached an accuracy of 87.69\%, the duration prediction on binned data was less successful and deviated by an average of 0.69 bins. A challenge for the location type classification, as well as for the subsequent components, was the imbalance of trips and connections as well as the low accuracy of the trajectory data. The imbalance is grounded in the fact that most users exhibit strong habitual patterns (e.g. home > work), while other patterns are rather rare by comparison. The accuracy problem derives from the energy-saving location sampling mode, which creates less accurate results. Those locations are then used to build a network that represents the user's spatio-temporal behaviour. An initial untrained ANN to predict movement on the network only reached 46\% average accuracy. Only lowering the number of included edges, focusing on more common trips, increased the performance. In order to further improve the algorithm, the spatial trajectories were introduced into the predictions. To overcome the accuracy problem, trips between locations were clustered into so-called spatial corridors, which were intersected with the user's current trajectory. The resulting intersected trips were ranked through a k-nearest-neighbour algorithm. This increased the performance to 56\%. In a final step, a combination of a network and spatial clustering algorithm was built in order to create clusters, therein reducing the variety of possible trips. By only predicting the destination cluster instead of the exact location, it is possible to increase the performance to 75\% including all classes. A final set of components shows in two exemplary ways how to deduce additional inferences from the underlying spatio-temporal data. The first example presents a novel concept for predicting the 'potential memorisation index' for a certain location. The index is based on a cognitive model which derives the index from the user's activity data in that area. The second example embeds each location in its urban fabric and thereby enriches its cluster's metadata by further describing the temporal-semantic activity in an area (e.g. going to restaurants at noon). The success of the client-side classification and prediction approach, despite the challenges of inaccurate and imbalanced data, supports the claimed benefits of the client-side modelling concept. Since modern data-driven services at some point do need to receive user data, the thesis' computational model concludes with a concept for applying generalisation to semantic, temporal, and spatial data before sharing it with the remote service in order to comply with the overall goal to improve data privacy. In this context, the potentials of ensemble training (in regards to ANNs) are discussed in order to highlight the potential of only sharing the trained ANN instead of the raw input data. While the results of our evaluation support the assets of the proposed framework, there are two important downsides of our approach compared to server-side modelling. First, both of these server-side advantages are rooted in the server's access to multiple users' data. This allows a remote service to predict spatio-in the user-specific data, which represents the second downside. While minor classes will likely be minor classes in a bigger dataset as well, for each class, there will still be more variety than in the user-specific dataset. The author emphasises that the approach presented in this work holds the potential to change the privacy paradigm in modern data-driven services. Finding combinations of client- and server-side modelling could prove a promising new path for data-driven innovation. Beyond the technological perspective, throughout the thesis the author also offers a critical view on the data- and technology-driven development of this work. By introducing the client-side modelling with user-specific artificial neural networks, users generate their own algorithm. Those user-specific algorithms are influenced less by generalised biases or developers' prejudices. Therefore, the user develops a more diverse and individual perspective through his or her user model. This concept picks up the idea of critical cartography, which questions the status quo of how space is perceived and represented.}, language = {en} } @phdthesis{Mientus2023, author = {Mientus, Lukas}, title = {Reflexion und Reflexivit{\"a}t}, doi = {10.25932/publishup-61000}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-610003}, school = {Universit{\"a}t Potsdam}, pages = {121}, year = {2023}, abstract = {Reflexion gilt in der Lehrkr{\"a}ftebildung als eine Schl{\"u}sselkategorie der professionellen Entwicklung. Entsprechend wird auf vielf{\"a}ltige Weise die Qualit{\"a}t reflexionsbezogener Kompetenzen untersucht. Eine Herausforderung hierbei kann in der Annahme bestehen, von der Analyse schriftlicher Reflexionen unmittelbar auf die Reflexivit{\"a}t einer Person zu schließen, da Reflexion stets kontextspezifisch als Abbild reflexionsbezogener Argumentationsprozesse angesehen werden sollte und reflexionsbezogenen Dispositionen unterliegt. Auch kann die Qualit{\"a}t einer Reflexion auf mehreren Dimensionen bewertet werden, ohne quantifizierbare, absolute Aussagen treffen zu k{\"o}nnen. Daher wurden im Rahmen einer Physik-Videovignette N = 134 schriftliche Fremdreflexionen verfasst und kontextspezifische reflexionsbezogene Dispositionen erhoben. Expert*innen erstellten theoriegeleitet Qualit{\"a}tsbewertungen zur Breite, Tiefe, Koh{\"a}renz und Spezifit{\"a}t eines jeden Reflexionstextes. Unter Verwendung computerbasierter Klassifikations- und Analyseverfahren wurden weitere Textmerkmale erhoben. Mittels explorativer Faktorenanalyse konnten die Faktoren Qualit{\"a}t, Quantit{\"a}t und Deskriptivit{\"a}t gefunden werden. Da alle konventionell eingesch{\"a}tzten Qualit{\"a}tsbewertungen durch einen Faktor repr{\"a}sentiert wurden, konnte ein maximales Qualit{\"a}tskorrelat kalkuliert werden, zu welchem jede schriftliche Fremdreflexion im Rahmen der vorliegenden Vignette eine computerbasiert bestimmbare Distanz aufweist. Diese Distanz zum maximalen Qualit{\"a}tskorrelat konnte validiert werden und kann die Qualit{\"a}t der schriftlichen Reflexionen unabh{\"a}ngig von menschlichen Ressourcen quantifiziert repr{\"a}sentieren. Abschließend konnte identifiziert werden, dass ausgew{\"a}hlte Dispositionen in unterschiedlichem Maße mit der Reflexionsqualit{\"a}t zusammenh{\"a}ngen. So konnten beispielsweise bezogen auf das Physik-Fachwissen minimale Zusammenh{\"a}nge identifiziert werden, wohingegen Werthaltung sowie wahrgenommene Unterrichtsqualit{\"a}t eng mit der Qualit{\"a}t einer schriftlichen Reflexion in Verbindung stehen k{\"o}nnen. Es wird geschlussfolgert, dass reflexionsbezogene Dispositionen moderierenden Einfluss auf Reflexionen nehmen k{\"o}nnen. Es wird empfohlen bei der Erhebung von Reflexion mit dem Ziel der Kompetenzmessung ausgew{\"a}hlte Dispositionen mit zu erheben. Weiter verdeutlicht diese Arbeit die M{\"o}glichkeit, aussagekr{\"a}ftige Quantifizierungen auch in der Analyse komplexer Konstrukte vorzunehmen. Durch computerbasierte Qualit{\"a}tsabsch{\"a}tzungen k{\"o}nnen objektive und individuelle Analysen und differenzierteres automatisiertes Feedback erm{\"o}glicht werden.}, language = {de} } @phdthesis{Najafi2023, author = {Najafi, Pejman}, title = {Leveraging data science \& engineering for advanced security operations}, doi = {10.25932/publishup-61225}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-612257}, school = {Universit{\"a}t Potsdam}, pages = {xix, 180}, year = {2023}, abstract = {The Security Operations Center (SOC) represents a specialized unit responsible for managing security within enterprises. To aid in its responsibilities, the SOC relies heavily on a Security Information and Event Management (SIEM) system that functions as a centralized repository for all security-related data, providing a comprehensive view of the organization's security posture. Due to the ability to offer such insights, SIEMS are considered indispensable tools facilitating SOC functions, such as monitoring, threat detection, and incident response. Despite advancements in big data architectures and analytics, most SIEMs fall short of keeping pace. Architecturally, they function merely as log search engines, lacking the support for distributed large-scale analytics. Analytically, they rely on rule-based correlation, neglecting the adoption of more advanced data science and machine learning techniques. This thesis first proposes a blueprint for next-generation SIEM systems that emphasize distributed processing and multi-layered storage to enable data mining at a big data scale. Next, with the architectural support, it introduces two data mining approaches for advanced threat detection as part of SOC operations. First, a novel graph mining technique that formulates threat detection within the SIEM system as a large-scale graph mining and inference problem, built on the principles of guilt-by-association and exempt-by-reputation. The approach entails the construction of a Heterogeneous Information Network (HIN) that models shared characteristics and associations among entities extracted from SIEM-related events/logs. Thereon, a novel graph-based inference algorithm is used to infer a node's maliciousness score based on its associations with other entities in the HIN. Second, an innovative outlier detection technique that imitates a SOC analyst's reasoning process to find anomalies/outliers. The approach emphasizes explainability and simplicity, achieved by combining the output of simple context-aware univariate submodels that calculate an outlier score for each entry. Both approaches were tested in academic and real-world settings, demonstrating high performance when compared to other algorithms as well as practicality alongside a large enterprise's SIEM system. This thesis establishes the foundation for next-generation SIEM systems that can enhance today's SOCs and facilitate the transition from human-centric to data-driven security operations.}, language = {en} } @misc{PanzerBenderGronau2022, author = {Panzer, Marcel and Bender, Benedict and Gronau, Norbert}, title = {Neural agent-based production planning and control}, series = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, journal = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, issn = {1867-5808}, doi = {10.25932/publishup-60477}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-604777}, pages = {26}, year = {2022}, abstract = {Nowadays, production planning and control must cope with mass customization, increased fluctuations in demand, and high competition pressures. Despite prevailing market risks, planning accuracy and increased adaptability in the event of disruptions or failures must be ensured, while simultaneously optimizing key process indicators. To manage that complex task, neural networks that can process large quantities of high-dimensional data in real time have been widely adopted in recent years. Although these are already extensively deployed in production systems, a systematic review of applications and implemented agent embeddings and architectures has not yet been conducted. The main contribution of this paper is to provide researchers and practitioners with an overview of applications and applied embeddings and to motivate further research in neural agent-based production. Findings indicate that neural agents are not only deployed in diverse applications, but are also increasingly implemented in multi-agent environments or in combination with conventional methods — leveraging performances compared to benchmarks and reducing dependence on human experience. This not only implies a more sophisticated focus on distributed production resources, but also broadening the perspective from a local to a global scale. Nevertheless, future research must further increase scalability and reproducibility to guarantee a simplified transfer of results to reality.}, language = {en} } @misc{PanzerBenderGronau2021, author = {Panzer, Marcel and Bender, Benedict and Gronau, Norbert}, title = {Deep reinforcement learning in production planning and control}, series = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, journal = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, issn = {2701-6277}, doi = {10.25932/publishup-60572}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-605722}, pages = {13}, year = {2021}, abstract = {Increasingly fast development cycles and individualized products pose major challenges for today's smart production systems in times of industry 4.0. The systems must be flexible and continuously adapt to changing conditions while still guaranteeing high throughputs and robustness against external disruptions. Deep reinforcement learning (RL) algorithms, which already reached impressive success with Google DeepMind's AlphaGo, are increasingly transferred to production systems to meet related requirements. Unlike supervised and unsupervised machine learning techniques, deep RL algorithms learn based on recently collected sensorand process-data in direct interaction with the environment and are able to perform decisions in real-time. As such, deep RL algorithms seem promising given their potential to provide decision support in complex environments, as production systems, and simultaneously adapt to changing circumstances. While different use-cases for deep RL emerged, a structured overview and integration of findings on their application are missing. To address this gap, this contribution provides a systematic literature review of existing deep RL applications in the field of production planning and control as well as production logistics. From a performance perspective, it became evident that deep RL can beat heuristics significantly in their overall performance and provides superior solutions to various industrial use-cases. Nevertheless, safety and reliability concerns must be overcome before the widespread use of deep RL is possible which presumes more intensive testing of deep RL in real world applications besides the already ongoing intensive simulations.}, language = {en} } @article{PanzerBenderGronau2022, author = {Panzer, Marcel and Bender, Benedict and Gronau, Norbert}, title = {Neural agent-based production planning and control}, series = {Journal of Manufacturing Systems}, volume = {65}, journal = {Journal of Manufacturing Systems}, publisher = {Elsevier}, address = {Amsterdam}, issn = {0278-6125}, doi = {10.1016/j.jmsy.2022.10.019}, pages = {743 -- 766}, year = {2022}, abstract = {Nowadays, production planning and control must cope with mass customization, increased fluctuations in demand, and high competition pressures. Despite prevailing market risks, planning accuracy and increased adaptability in the event of disruptions or failures must be ensured, while simultaneously optimizing key process indicators. To manage that complex task, neural networks that can process large quantities of high-dimensional data in real time have been widely adopted in recent years. Although these are already extensively deployed in production systems, a systematic review of applications and implemented agent embeddings and architectures has not yet been conducted. The main contribution of this paper is to provide researchers and practitioners with an overview of applications and applied embeddings and to motivate further research in neural agent-based production. Findings indicate that neural agents are not only deployed in diverse applications, but are also increasingly implemented in multi-agent environments or in combination with conventional methods — leveraging performances compared to benchmarks and reducing dependence on human experience. This not only implies a more sophisticated focus on distributed production resources, but also broadening the perspective from a local to a global scale. Nevertheless, future research must further increase scalability and reproducibility to guarantee a simplified transfer of results to reality.}, language = {en} } @inproceedings{PanzerBenderGronau2021, author = {Panzer, Marcel and Bender, Benedict and Gronau, Norbert}, title = {Deep reinforcement learning in production planning and control}, series = {Proceedings of the Conference on Production Systems and Logistics}, booktitle = {Proceedings of the Conference on Production Systems and Logistics}, publisher = {publish-Ing.}, address = {Hannover}, issn = {2701-6277}, doi = {10.15488/11238}, pages = {535 -- 545}, year = {2021}, abstract = {Increasingly fast development cycles and individualized products pose major challenges for today's smart production systems in times of industry 4.0. The systems must be flexible and continuously adapt to changing conditions while still guaranteeing high throughputs and robustness against external disruptions. Deep rein- forcement learning (RL) algorithms, which already reached impressive success with Google DeepMind's AlphaGo, are increasingly transferred to production systems to meet related requirements. Unlike supervised and unsupervised machine learning techniques, deep RL algorithms learn based on recently collected sensor- and process-data in direct interaction with the environment and are able to perform decisions in real-time. As such, deep RL algorithms seem promising given their potential to provide decision support in complex environments, as production systems, and simultaneously adapt to changing circumstances. While different use-cases for deep RL emerged, a structured overview and integration of findings on their application are missing. To address this gap, this contribution provides a systematic literature review of existing deep RL applications in the field of production planning and control as well as production logistics. From a performance perspective, it became evident that deep RL can beat heuristics significantly in their overall performance and provides superior solutions to various industrial use-cases. Nevertheless, safety and reliability concerns must be overcome before the widespread use of deep RL is possible which presumes more intensive testing of deep RL in real world applications besides the already ongoing intensive simulations.}, language = {en} } @unpublished{PrasseGrubenMachlikaetal.2016, author = {Prasse, Paul and Gruben, Gerrit and Machlika, Lukas and Pevny, Tomas and Sofka, Michal and Scheffer, Tobias}, title = {Malware Detection by HTTPS Traffic Analysis}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-100942}, pages = {10}, year = {2016}, abstract = {In order to evade detection by network-traffic analysis, a growing proportion of malware uses the encrypted HTTPS protocol. We explore the problem of detecting malware on client computers based on HTTPS traffic analysis. In this setting, malware has to be detected based on the host IP address, ports, timestamp, and data volume information of TCP/IP packets that are sent and received by all the applications on the client. We develop a scalable protocol that allows us to collect network flows of known malicious and benign applications as training data and derive a malware-detection method based on a neural networks and sequence classification. We study the method's ability to detect known and new, unknown malware in a large-scale empirical study.}, language = {en} } @book{RanaMohapatraSidorovaetal.2022, author = {Rana, Kaushik and Mohapatra, Durga Prasad and Sidorova, Julia and Lundberg, Lars and Sk{\"o}ld, Lars and Lopes Grim, Lu{\´i}s Fernando and Sampaio Gradvohl, Andr{\´e} Leon and Cremerius, Jonas and Siegert, Simon and Weltzien, Anton von and Baldi, Annika and Klessascheck, Finn and Kalancha, Svitlana and Lichtenstein, Tom and Shaabani, Nuhad and Meinel, Christoph and Friedrich, Tobias and Lenzner, Pascal and Schumann, David and Wiese, Ingmar and Sarna, Nicole and Wiese, Lena and Tashkandi, Araek Sami and van der Walt, Est{\´e}e and Eloff, Jan H. P. and Schmidt, Christopher and H{\"u}gle, Johannes and Horschig, Siegfried and Uflacker, Matthias and Najafi, Pejman and Sapegin, Andrey and Cheng, Feng and Stojanovic, Dragan and Stojnev Ilić, Aleksandra and Djordjevic, Igor and Stojanovic, Natalija and Predic, Bratislav and Gonz{\´a}lez-Jim{\´e}nez, Mario and de Lara, Juan and Mischkewitz, Sven and Kainz, Bernhard and van Hoorn, Andr{\´e} and Ferme, Vincenzo and Schulz, Henning and Knigge, Marlene and Hecht, Sonja and Prifti, Loina and Krcmar, Helmut and Fabian, Benjamin and Ermakova, Tatiana and Kelkel, Stefan and Baumann, Annika and Morgenstern, Laura and Plauth, Max and Eberhard, Felix and Wolff, Felix and Polze, Andreas and Cech, Tim and Danz, Noel and Noack, Nele Sina and Pirl, Lukas and Beilharz, Jossekin Jakob and De Oliveira, Roberto C. L. and Soares, F{\´a}bio Mendes and Juiz, Carlos and Bermejo, Belen and M{\"u}hle, Alexander and Gr{\"u}ner, Andreas and Saxena, Vageesh and Gayvoronskaya, Tatiana and Weyand, Christopher and Krause, Mirko and Frank, Markus and Bischoff, Sebastian and Behrens, Freya and R{\"u}ckin, Julius and Ziegler, Adrian and Vogel, Thomas and Tran, Chinh and Moser, Irene and Grunske, Lars and Sz{\´a}rnyas, G{\´a}bor and Marton, J{\´o}zsef and Maginecz, J{\´a}nos and Varr{\´o}, D{\´a}niel and Antal, J{\´a}nos Benjamin}, title = {HPI Future SOC Lab - Proceedings 2018}, number = {151}, editor = {Meinel, Christoph and Polze, Andreas and Beins, Karsten and Strotmann, Rolf and Seibold, Ulrich and R{\"o}dszus, Kurt and M{\"u}ller, J{\"u}rgen}, publisher = {Universit{\"a}tsverlag Potsdam}, address = {Potsdam}, isbn = {978-3-86956-547-7}, issn = {1613-5652}, doi = {10.25932/publishup-56371}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-563712}, publisher = {Universit{\"a}t Potsdam}, pages = {x, 277}, year = {2022}, abstract = {The "HPI Future SOC Lab" is a cooperation of the Hasso Plattner Institute (HPI) and industry partners. Its mission is to enable and promote exchange and interaction between the research community and the industry partners. The HPI Future SOC Lab provides researchers with free of charge access to a complete infrastructure of state of the art hard and software. This infrastructure includes components, which might be too expensive for an ordinary research environment, such as servers with up to 64 cores and 2 TB main memory. The offerings address researchers particularly from but not limited to the areas of computer science and business information systems. Main areas of research include cloud computing, parallelization, and In-Memory technologies. This technical report presents results of research projects executed in 2018. Selected projects have presented their results on April 17th and November 14th 2017 at the Future SOC Lab Day events.}, language = {en} } @phdthesis{Rezaei2019, author = {Rezaei, Mina}, title = {Deep representation learning from imbalanced medical imaging}, doi = {10.25932/publishup-44275}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-442759}, school = {Universit{\"a}t Potsdam}, pages = {xxviii, 160}, year = {2019}, abstract = {Medical imaging plays an important role in disease diagnosis, treatment planning, and clinical monitoring. One of the major challenges in medical image analysis is imbalanced training data, in which the class of interest is much rarer than the other classes. Canonical machine learning algorithms suppose that the number of samples from different classes in the training dataset is roughly similar or balance. Training a machine learning model on an imbalanced dataset can introduce unique challenges to the learning problem. A model learned from imbalanced training data is biased towards the high-frequency samples. The predicted results of such networks have low sensitivity and high precision. In medical applications, the cost of misclassification of the minority class could be more than the cost of misclassification of the majority class. For example, the risk of not detecting a tumor could be much higher than referring to a healthy subject to a doctor. The current Ph.D. thesis introduces several deep learning-based approaches for handling class imbalanced problems for learning multi-task such as disease classification and semantic segmentation. At the data-level, the objective is to balance the data distribution through re-sampling the data space: we propose novel approaches to correct internal bias towards fewer frequency samples. These approaches include patient-wise batch sampling, complimentary labels, supervised and unsupervised minority oversampling using generative adversarial networks for all. On the other hand, at algorithm-level, we modify the learning algorithm to alleviate the bias towards majority classes. In this regard, we propose different generative adversarial networks for cost-sensitive learning, ensemble learning, and mutual learning to deal with highly imbalanced imaging data. We show evidence that the proposed approaches are applicable to different types of medical images of varied sizes on different applications of routine clinical tasks, such as disease classification and semantic segmentation. Our various implemented algorithms have shown outstanding results on different medical imaging challenges.}, language = {en} } @phdthesis{Risch2020, author = {Risch, Julian}, title = {Reader comment analysis on online news platforms}, doi = {10.25932/publishup-48922}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-489222}, school = {Universit{\"a}t Potsdam}, pages = {xi, 135}, year = {2020}, abstract = {Comment sections of online news platforms are an essential space to express opinions and discuss political topics. However, the misuse by spammers, haters, and trolls raises doubts about whether the benefits justify the costs of the time-consuming content moderation. As a consequence, many platforms limited or even shut down comment sections completely. In this thesis, we present deep learning approaches for comment classification, recommendation, and prediction to foster respectful and engaging online discussions. The main focus is on two kinds of comments: toxic comments, which make readers leave a discussion, and engaging comments, which make readers join a discussion. First, we discourage and remove toxic comments, e.g., insults or threats. To this end, we present a semi-automatic comment moderation process, which is based on fine-grained text classification models and supports moderators. Our experiments demonstrate that data augmentation, transfer learning, and ensemble learning allow training robust classifiers even on small datasets. To establish trust in the machine-learned models, we reveal which input features are decisive for their output with attribution-based explanation methods. Second, we encourage and highlight engaging comments, e.g., serious questions or factual statements. We automatically identify the most engaging comments, so that readers need not scroll through thousands of comments to find them. The model training process builds on upvotes and replies as a measure of reader engagement. We also identify comments that address the article authors or are otherwise relevant to them to support interactions between journalists and their readership. Taking into account the readers' interests, we further provide personalized recommendations of discussions that align with their favored topics or involve frequent co-commenters. Our models outperform multiple baselines and recent related work in experiments on comment datasets from different platforms.}, language = {en} } @article{RyoJeschkeRilligetal.2020, author = {Ryo, Masahiro and Jeschke, Jonathan M. and Rillig, Matthias C. and Heger, Tina}, title = {Machine learning with the hierarchy-of-hypotheses (HoH) approach discovers novel pattern in studies on biological invasions}, series = {Research synthesis methods}, volume = {11}, journal = {Research synthesis methods}, number = {1}, publisher = {Wiley}, address = {Hoboken}, issn = {1759-2879}, doi = {10.1002/jrsm.1363}, pages = {66 -- 73}, year = {2020}, abstract = {Research synthesis on simple yet general hypotheses and ideas is challenging in scientific disciplines studying highly context-dependent systems such as medical, social, and biological sciences. This study shows that machine learning, equation-free statistical modeling of artificial intelligence, is a promising synthesis tool for discovering novel patterns and the source of controversy in a general hypothesis. We apply a decision tree algorithm, assuming that evidence from various contexts can be adequately integrated in a hierarchically nested structure. As a case study, we analyzed 163 articles that studied a prominent hypothesis in invasion biology, the enemy release hypothesis. We explored if any of the nine attributes that classify each study can differentiate conclusions as classification problem. Results corroborated that machine learning can be useful for research synthesis, as the algorithm could detect patterns that had been already focused in previous narrative reviews. Compared with the previous synthesis study that assessed the same evidence collection based on experts' judgement, the algorithm has newly proposed that the studies focusing on Asian regions mostly supported the hypothesis, suggesting that more detailed investigations in these regions can enhance our understanding of the hypothesis. We suggest that machine learning algorithms can be a promising synthesis tool especially where studies (a) reformulate a general hypothesis from different perspectives, (b) use different methods or variables, or (c) report insufficient information for conducting meta-analyses.}, language = {en} } @misc{RyoJeschkeRilligetal.2020, author = {Ryo, Masahiro and Jeschke, Jonathan M. and Rillig, Matthias C. and Heger, Tina}, title = {Machine learning with the hierarchy-of-hypotheses (HoH) approach discovers novel pattern in studies on biological invasions}, series = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {1171}, issn = {1866-8372}, doi = {10.25932/publishup-51764}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-517643}, pages = {66 -- 73}, year = {2020}, abstract = {Research synthesis on simple yet general hypotheses and ideas is challenging in scientific disciplines studying highly context-dependent systems such as medical, social, and biological sciences. This study shows that machine learning, equation-free statistical modeling of artificial intelligence, is a promising synthesis tool for discovering novel patterns and the source of controversy in a general hypothesis. We apply a decision tree algorithm, assuming that evidence from various contexts can be adequately integrated in a hierarchically nested structure. As a case study, we analyzed 163 articles that studied a prominent hypothesis in invasion biology, the enemy release hypothesis. We explored if any of the nine attributes that classify each study can differentiate conclusions as classification problem. Results corroborated that machine learning can be useful for research synthesis, as the algorithm could detect patterns that had been already focused in previous narrative reviews. Compared with the previous synthesis study that assessed the same evidence collection based on experts' judgement, the algorithm has newly proposed that the studies focusing on Asian regions mostly supported the hypothesis, suggesting that more detailed investigations in these regions can enhance our understanding of the hypothesis. We suggest that machine learning algorithms can be a promising synthesis tool especially where studies (a) reformulate a general hypothesis from different perspectives, (b) use different methods or variables, or (c) report insufficient information for conducting meta-analyses.}, language = {en} } @phdthesis{Sapegin2018, author = {Sapegin, Andrey}, title = {High-Speed Security Log Analytics Using Hybrid Outlier Detection}, doi = {10.25932/publishup-42611}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-426118}, school = {Universit{\"a}t Potsdam}, pages = {162}, year = {2018}, abstract = {The rapid development and integration of Information Technologies over the last decades influenced all areas of our life, including the business world. Yet not only the modern enterprises become digitalised, but also security and criminal threats move into the digital sphere. To withstand these threats, modern companies must be aware of all activities within their computer networks. The keystone for such continuous security monitoring is a Security Information and Event Management (SIEM) system that collects and processes all security-related log messages from the entire enterprise network. However, digital transformations and technologies, such as network virtualisation and widespread usage of mobile communications, lead to a constantly increasing number of monitored devices and systems. As a result, the amount of data that has to be processed by a SIEM system is increasing rapidly. Besides that, in-depth security analysis of the captured data requires the application of rather sophisticated outlier detection algorithms that have a high computational complexity. Existing outlier detection methods often suffer from performance issues and are not directly applicable for high-speed and high-volume analysis of heterogeneous security-related events, which becomes a major challenge for modern SIEM systems nowadays. This thesis provides a number of solutions for the mentioned challenges. First, it proposes a new SIEM system architecture for high-speed processing of security events, implementing parallel, in-memory and in-database processing principles. The proposed architecture also utilises the most efficient log format for high-speed data normalisation. Next, the thesis offers several novel high-speed outlier detection methods, including generic Hybrid Outlier Detection that can efficiently be used for Big Data analysis. Finally, the special User Behaviour Outlier Detection is proposed for better threat detection and analysis of particular user behaviour cases. The proposed architecture and methods were evaluated in terms of both performance and accuracy, as well as compared with classical architecture and existing algorithms. These evaluations were performed on multiple data sets, including simulated data, well-known public intrusion detection data set, and real data from the large multinational enterprise. The evaluation results have proved the high performance and efficacy of the developed methods. All concepts proposed in this thesis were integrated into the prototype of the SIEM system, capable of high-speed analysis of Big Security Data, which makes this integrated SIEM platform highly relevant for modern enterprise security applications.}, language = {en} } @article{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Hesse, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference}, series = {Water resources research}, volume = {56}, journal = {Water resources research}, number = {5}, publisher = {American Geophysical Union}, address = {Washington}, issn = {0043-1397}, doi = {10.1029/2019WR025924}, pages = {10}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @misc{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Heße, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference: a case study for flooding events across Germany}, series = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {5}, issn = {1866-8372}, doi = {10.25932/publishup-52384}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-523843}, pages = {12}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @article{SchmidtHesseAttingeretal.2020, author = {Schmidt, Lennart and Heße, Falk and Attinger, Sabine and Kumar, Rohini}, title = {Challenges in applying machine learning models for hydrological inference: a case study for flooding events across Germany}, series = {Water Resources Research}, volume = {56}, journal = {Water Resources Research}, number = {5}, publisher = {John Wiley \& Sons, Inc.}, address = {New Jersey}, pages = {10}, year = {2020}, abstract = {Machine learning (ML) algorithms are being increasingly used in Earth and Environmental modeling studies owing to the ever-increasing availability of diverse data sets and computational resources as well as advancement in ML algorithms. Despite advances in their predictive accuracy, the usefulness of ML algorithms for inference remains elusive. In this study, we employ two popular ML algorithms, artificial neural networks and random forest, to analyze a large data set of flood events across Germany with the goals to analyze their predictive accuracy and their usability to provide insights to hydrologic system functioning. The results of the ML algorithms are contrasted against a parametric approach based on multiple linear regression. For analysis, we employ a model-agnostic framework named Permuted Feature Importance to derive the influence of models' predictors. This allows us to compare the results of different algorithms for the first time in the context of hydrology. Our main findings are that (1) the ML models achieve higher prediction accuracy than linear regression, (2) the results reflect basic hydrological principles, but (3) further inference is hindered by the heterogeneity of results across algorithms. Thus, we conclude that the problem of equifinality as known from classical hydrological modeling also exists for ML and severely hampers its potential for inference. To account for the observed problems, we propose that when employing ML for inference, this should be made by using multiple algorithms and multiple methods, of which the latter should be embedded in a cross-validation routine.}, language = {en} } @phdthesis{Schroeter2020, author = {Schr{\"o}ter, Kai}, title = {Improved flood risk assessment}, doi = {10.25932/publishup-48024}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-480240}, school = {Universit{\"a}t Potsdam}, pages = {408}, year = {2020}, abstract = {Rivers have always flooded their floodplains. Over 2.5 billion people worldwide have been affected by flooding in recent decades. The economic damage is also considerable, averaging 100 billion US dollars per year. There is no doubt that damage and other negative effects of floods can be avoided. However, this has a price: financially and politically. Costs and benefits can be estimated through risk assessments. Questions about the location and frequency of floods, about the objects that could be affected and their vulnerability are of importance for flood risk managers, insurance companies and politicians. Thus, both variables and factors from the fields of hydrology and sociol-economics play a role with multi-layered connections. One example are dikes along a river, which on the one hand contain floods, but on the other hand, by narrowing the natural floodplains, accelerate the flood discharge and increase the danger of flooding for the residents downstream. Such larger connections must be included in the assessment of flood risk. However, in current procedures this is accompanied by simplifying assumptions. Risk assessments are therefore fuzzy and associated with uncertainties. This thesis investigates the benefits and possibilities of new data sources for improving flood risk assessment. New methods and models are developed, which take the mentioned interrelations better into account and also quantify the existing uncertainties of the model results, and thus enable statements about the reliability of risk estimates. For this purpose, data on flood events from various sources are collected and evaluated. This includes precipitation and flow records at measuring stations as well as for instance images from social media, which can help to delineate the flooded areas and estimate flood damage with location information. Machine learning methods have been successfully used to recognize and understand correlations between floods and impacts from a wide range of data and to develop improved models. Risk models help to develop and evaluate strategies to reduce flood risk. These tools also provide advanced insights into the interplay of various factors and on the expected consequences of flooding. This work shows progress in terms of an improved assessment of flood risks by using diverse data from different sources with innovative methods as well as by the further development of models. Flood risk is variable due to economic and climatic changes, and other drivers of risk. In order to keep the knowledge about flood risks up-to-date, robust, efficient and adaptable methods as proposed in this thesis are of increasing importance.}, language = {en} } @article{SchudomaLarhlimiWalther2011, author = {Schudoma, Christian and Larhlimi, Abdelhalim and Walther, Dirk}, title = {The influence of the local sequence environment on RNA loop structures}, series = {RNA : a publication of the RNA Society}, volume = {17}, journal = {RNA : a publication of the RNA Society}, number = {7}, publisher = {Cold Spring Harbor Laboratory Press}, address = {Cold Spring Harbor, NY}, issn = {1355-8382}, doi = {10.1261/rna.2550211}, pages = {1247 -- 1257}, year = {2011}, abstract = {RNA folding is assumed to be a hierarchical process. The secondary structure of an RNA molecule, signified by base-pairing and stacking interactions between the paired bases, is formed first. Subsequently, the RNA molecule adopts an energetically favorable three-dimensional conformation in the structural space determined mainly by the rotational degrees of freedom associated with the backbone of regions of unpaired nucleotides (loops). To what extent the backbone conformation of RNA loops also results from interactions within the local sequence context or rather follows global optimization constraints alone has not been addressed yet. Because the majority of base stacking interactions are exerted locally, a critical influence of local sequence on local structure appears plausible. Thus, local loop structure ought to be predictable, at least in part, from the local sequence context alone. To test this hypothesis, we used Random Forests on a nonredundant data set of unpaired nucleotides extracted from 97 X-ray structures from the Protein Data Bank (PDB) to predict discrete backbone angle conformations given by the discretized eta/theta-pseudo-torsional space. Predictions on balanced sets with four to six conformational classes using local sequence information yielded average accuracies of up to 55\%, thus significantly better than expected by chance (17\%-25\%). Bases close to the central nucleotide appear to be most tightly linked to its conformation. Our results suggest that RNA loop structure does not only depend on long-range base-pairing interactions; instead, it appears that local sequence context exerts a significant influence on the formation of the local loop structure.}, language = {en} } @article{SeewannVerwiebeBuderetal.2022, author = {Seewann, Lena and Verwiebe, Roland and Buder, Claudia and Fritsch, Nina-Sophie}, title = {"Broadcast your gender."}, series = {Frontiers in Big Data}, journal = {Frontiers in Big Data}, number = {5}, publisher = {Frontiers}, address = {Lausanne, Schweiz}, issn = {2624-909X}, doi = {10.3389/fdata.2022.908636}, pages = {16}, year = {2022}, abstract = {Social media platforms provide a large array of behavioral data relevant to social scientific research. However, key information such as sociodemographic characteristics of agents are often missing. This paper aims to compare four methods of classifying social attributes from text. Specifically, we are interested in estimating the gender of German social media creators. By using the example of a random sample of 200 YouTube channels, we compare several classification methods, namely (1) a survey among university staff, (2) a name dictionary method with the World Gender Name Dictionary as a reference list, (3) an algorithmic approach using the website gender-api.com, and (4) a Multinomial Na{\"i}ve Bayes (MNB) machine learning technique. These different methods identify gender attributes based on YouTube channel names and descriptions in German but are adaptable to other languages. Our contribution will evaluate the share of identifiable channels, accuracy and meaningfulness of classification, as well as limits and benefits of each approach. We aim to address methodological challenges connected to classifying gender attributes for YouTube channels as well as related to reinforcing stereotypes and ethical implications.}, language = {en} } @misc{SeewannVerwiebeBuderetal.2022, author = {Seewann, Lena and Verwiebe, Roland and Buder, Claudia and Fritsch, Nina-Sophie}, title = {"Broadcast your gender."}, series = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, journal = {Zweitver{\"o}ffentlichungen der Universit{\"a}t Potsdam : Wirtschafts- und Sozialwissenschaftliche Reihe}, number = {152}, issn = {1867-5808}, doi = {10.25932/publishup-56628}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-566287}, pages = {16}, year = {2022}, abstract = {Social media platforms provide a large array of behavioral data relevant to social scientific research. However, key information such as sociodemographic characteristics of agents are often missing. This paper aims to compare four methods of classifying social attributes from text. Specifically, we are interested in estimating the gender of German social media creators. By using the example of a random sample of 200 YouTube channels, we compare several classification methods, namely (1) a survey among university staff, (2) a name dictionary method with the World Gender Name Dictionary as a reference list, (3) an algorithmic approach using the website gender-api.com, and (4) a Multinomial Na{\"i}ve Bayes (MNB) machine learning technique. These different methods identify gender attributes based on YouTube channel names and descriptions in German but are adaptable to other languages. Our contribution will evaluate the share of identifiable channels, accuracy and meaningfulness of classification, as well as limits and benefits of each approach. We aim to address methodological challenges connected to classifying gender attributes for YouTube channels as well as related to reinforcing stereotypes and ethical implications.}, language = {en} } @phdthesis{Seleem2023, author = {Seleem, Omar}, title = {Towards urban pluvial flood mapping using data-driven models}, doi = {10.25932/publishup-59813}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-598137}, school = {Universit{\"a}t Potsdam}, pages = {xv, 95}, year = {2023}, abstract = {Casualties and damages from urban pluvial flooding are increasing. Triggered by short, localized, and intensive rainfall events, urban pluvial floods can occur anywhere, even in areas without a history of flooding. Urban pluvial floods have relatively small temporal and spatial scales. Although cumulative losses from urban pluvial floods are comparable, most flood risk management and mitigation strategies focus on fluvial and coastal flooding. Numerical-physical-hydrodynamic models are considered the best tool to represent the complex nature of urban pluvial floods; however, they are computationally expensive and time-consuming. These sophisticated models make large-scale analysis and operational forecasting prohibitive. Therefore, it is crucial to evaluate and benchmark the performance of other alternative methods. The findings of this cumulative thesis are represented in three research articles. The first study evaluates two topographic-based methods to map urban pluvial flooding, fill-spill-merge (FSM) and topographic wetness index (TWI), by comparing them against a sophisticated hydrodynamic model. The FSM method identifies flood-prone areas within topographic depressions while the TWI method employs maximum likelihood estimation to calibrate a TWI threshold (τ) based on inundation maps from the 2D hydrodynamic model. The results point out that the FSM method outperforms the TWI method. The study highlights then the advantage and limitations of both methods. Data-driven models provide a promising alternative to computationally expensive hydrodynamic models. However, the literature lacks benchmarking studies to evaluate the different models' performance, advantages and limitations. Model transferability in space is a crucial problem. Most studies focus on river flooding, likely due to the relative availability of flow and rain gauge records for training and validation. Furthermore, they consider these models as black boxes. The second study uses a flood inventory for the city of Berlin and 11 predictive features which potentially indicate an increased pluvial flooding hazard to map urban pluvial flood susceptibility using a convolutional neural network (CNN), an artificial neural network (ANN) and the benchmarking machine learning models random forest (RF) and support vector machine (SVM). I investigate the influence of spatial resolution on the implemented models, the models' transferability in space and the importance of the predictive features. The results show that all models perform well and the RF models are superior to the other models within and outside the training domain. The models developed using fine spatial resolution (2 and 5 m) could better identify flood-prone areas. Finally, the results point out that aspect is the most important predictive feature for the CNN models, and altitude is for the other models. While flood susceptibility maps identify flood-prone areas, they do not represent flood variables such as velocity and depth which are necessary for effective flood risk management. To address this, the third study investigates data-driven models' transferability to predict urban pluvial floodwater depth and the models' ability to enhance their predictions using transfer learning techniques. It compares the performance of RF (the best-performing model in the previous study) and CNN models using 12 predictive features and output from a hydrodynamic model. The findings in the third study suggest that while CNN models tend to generalise and smooth the target function on the training dataset, RF models suffer from overfitting. Hence, RF models are superior for predictions inside the training domains but fail outside them while CNN models could control the relative loss in performance outside the training domains. Finally, the CNN models benefit more from transfer learning techniques than RF models, boosting their performance outside training domains. In conclusion, this thesis has evaluated both topographic-based methods and data-driven models to map urban pluvial flooding. However, further studies are crucial to have methods that completely overcome the limitation of 2D hydrodynamic models.}, language = {en} } @phdthesis{Sidarenka2019, author = {Sidarenka, Uladzimir}, title = {Sentiment analysis of German Twitter}, doi = {10.25932/publishup-43742}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-437422}, school = {Universit{\"a}t Potsdam}, pages = {vii, 217}, year = {2019}, abstract = {The immense popularity of online communication services in the last decade has not only upended our lives (with news spreading like wildfire on the Web, presidents announcing their decisions on Twitter, and the outcome of political elections being determined on Facebook) but also dramatically increased the amount of data exchanged on these platforms. Therefore, if we wish to understand the needs of modern society better and want to protect it from new threats, we urgently need more robust, higher-quality natural language processing (NLP) applications that can recognize such necessities and menaces automatically, by analyzing uncensored texts. Unfortunately, most NLP programs today have been created for standard language, as we know it from newspapers, or, in the best case, adapted to the specifics of English social media. This thesis reduces the existing deficit by entering the new frontier of German online communication and addressing one of its most prolific forms—users' conversations on Twitter. In particular, it explores the ways and means by how people express their opinions on this service, examines current approaches to automatic mining of these feelings, and proposes novel methods, which outperform state-of-the-art techniques. For this purpose, I introduce a new corpus of German tweets that have been manually annotated with sentiments, their targets and holders, as well as lexical polarity items and their contextual modifiers. Using these data, I explore four major areas of sentiment research: (i) generation of sentiment lexicons, (ii) fine-grained opinion mining, (iii) message-level polarity classification, and (iv) discourse-aware sentiment analysis. In the first task, I compare three popular groups of lexicon generation methods: dictionary-, corpus-, and word-embedding-based ones, finding that dictionary-based systems generally yield better polarity lists than the last two groups. Apart from this, I propose a linear projection algorithm, whose results surpass many existing automatically-generated lexicons. Afterwords, in the second task, I examine two common approaches to automatic prediction of sentiment spans, their sources, and targets: conditional random fields (CRFs) and recurrent neural networks, obtaining higher scores with the former model and improving these results even further by redefining the structure of CRF graphs. When dealing with message-level polarity classification, I juxtapose three major sentiment paradigms: lexicon-, machine-learning-, and deep-learning-based systems, and try to unite the first and last of these method groups by introducing a bidirectional neural network with lexicon-based attention. Finally, in order to make the new classifier aware of microblogs' discourse structure, I let it separately analyze the elementary discourse units of each tweet and infer the overall polarity of a message from the scores of its EDUs with the help of two new approaches: latent-marginalized CRFs and Recursive Dirichlet Process.}, language = {en} } @phdthesis{Smirnov2023, author = {Smirnov, Artem}, title = {Understanding the dynamics of the near-earth space environment utilizing long-term satellite observations}, doi = {10.25932/publishup-61371}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-613711}, school = {Universit{\"a}t Potsdam}, pages = {xxxvi, 286}, year = {2023}, abstract = {The near-Earth space environment is a highly complex system comprised of several regions and particle populations hazardous to satellite operations. The trapped particles in the radiation belts and ring current can cause significant damage to satellites during space weather events, due to deep dielectric and surface charging. Closer to Earth is another important region, the ionosphere, which delays the propagation of radio signals and can adversely affect navigation and positioning. In response to fluctuations in solar and geomagnetic activity, both the inner-magnetospheric and ionospheric populations can undergo drastic and sudden changes within minutes to hours, which creates a challenge for predicting their behavior. Given the increasing reliance of our society on satellite technology, improving our understanding and modeling of these populations is a matter of paramount importance. In recent years, numerous spacecraft have been launched to study the dynamics of particle populations in the near-Earth space, transforming it into a data-rich environment. To extract valuable insights from the abundance of available observations, it is crucial to employ advanced modeling techniques, and machine learning methods are among the most powerful approaches available. This dissertation employs long-term satellite observations to analyze the processes that drive particle dynamics, and builds interdisciplinary links between space physics and machine learning by developing new state-of-the-art models of the inner-magnetospheric and ionospheric particle dynamics. The first aim of this thesis is to investigate the behavior of electrons in Earth's radiation belts and ring current. Using ~18 years of electron flux observations from the Global Positioning System (GPS), we developed the first machine learning model of hundreds-of-keV electron flux at Medium Earth Orbit (MEO) that is driven solely by solar wind and geomagnetic indices and does not require auxiliary flux measurements as inputs. We then proceeded to analyze the directional distributions of electrons, and for the first time, used Fourier sine series to fit electron pitch angle distributions (PADs) in Earth's inner magnetosphere. We performed a superposed epoch analysis of 129 geomagnetic storms during the Van Allen Probes era and demonstrated that electron PADs have a strong energy-dependent response to geomagnetic activity. Additionally, we showed that the solar wind dynamic pressure could be used as a good predictor of the PAD dynamics. Using the observed dependencies, we created the first PAD model with a continuous dependence on L, magnetic local time (MLT) and activity, and developed two techniques to reconstruct near-equatorial electron flux observations from low-PA data using this model. The second objective of this thesis is to develop a novel model of the topside ionosphere. To achieve this goal, we collected observations from five of the most widely used ionospheric missions and intercalibrated these data sets. This allowed us to use these data jointly for model development, validation, and comparison with other existing empirical models. We demonstrated, for the first time, that ion density observations by Swarm Langmuir Probes exhibit overestimation (up to ~40-50\%) at low and mid-latitudes on the night side, and suggested that the influence of light ions could be a potential cause of this overestimation. To develop the topside model, we used 19 years of radio occultation (RO) electron density profiles, which were fitted with a Chapman function with a linear dependence of scale height on altitude. This approximation yields 4 parameters, namely the peak density and height of the F2-layer and the slope and intercept of the linear scale height trend, which were modeled using feedforward neural networks (NNs). The model was extensively validated against both RO and in-situ observations and was found to outperform the International Reference Ionosphere (IRI) model by up to an order of magnitude. Our analysis showed that the most substantial deviations of the IRI model from the data occur at altitudes of 100-200 km above the F2-layer peak. The developed NN-based ionospheric model reproduces the effects of various physical mechanisms observed in the topside ionosphere and provides highly accurate electron density predictions. This dissertation provides an extensive study of geospace dynamics, and the main results of this work contribute to the improvement of models of plasma populations in the near-Earth space environment.}, language = {en} } @article{SmirnovBerrendorfShpritsetal.2020, author = {Smirnov, Artem and Berrendorf, Max and Shprits, Yuri Y. and Kronberg, Elena A. and Allison, Hayley J. and Aseev, Nikita and Zhelavskaya, Irina and Morley, Steven K. and Reeves, Geoffrey D. and Carver, Matthew R. and Effenberger, Frederic}, title = {Medium energy electron flux in earth's outer radiation belt (MERLIN)}, series = {Space weather : the international journal of research and applications}, volume = {18}, journal = {Space weather : the international journal of research and applications}, number = {11}, publisher = {American geophysical union, AGU}, address = {Washington}, issn = {1542-7390}, doi = {10.1029/2020SW002532}, pages = {20}, year = {2020}, abstract = {The radiation belts of the Earth, filled with energetic electrons, comprise complex and dynamic systems that pose a significant threat to satellite operation. While various models of electron flux both for low and relativistic energies have been developed, the behavior of medium energy (120-600 keV) electrons, especially in the MEO region, remains poorly quantified. At these energies, electrons are driven by both convective and diffusive transport, and their prediction usually requires sophisticated 4D modeling codes. In this paper, we present an alternative approach using the Light Gradient Boosting (LightGBM) machine learning algorithm. The Medium Energy electRon fLux In Earth's outer radiatioN belt (MERLIN) model takes as input the satellite position, a combination of geomagnetic indices and solar wind parameters including the time history of velocity, and does not use persistence. MERLIN is trained on >15 years of the GPS electron flux data and tested on more than 1.5 years of measurements. Tenfold cross validation yields that the model predicts the MEO radiation environment well, both in terms of dynamics and amplitudes o f flux. Evaluation on the test set shows high correlation between the predicted and observed electron flux (0.8) and low values of absolute error. The MERLIN model can have wide space weather applications, providing information for the scientific community in the form of radiation belts reconstructions, as well as industry for satellite mission design, nowcast of the MEO environment, and surface charging analysis.}, language = {en} } @article{SprengerErbanSeddigetal.2017, author = {Sprenger, Heike and Erban, Alexander and Seddig, Sylvia and Rudack, Katharina and Thalhammer, Anja and Le, Mai Q. and Walther, Dirk and Zuther, Ellen and Koehl, Karin I. and Kopka, Joachim and Hincha, Dirk K.}, title = {Metabolite and transcript markers for the prediction of potato drought tolerance}, series = {Plant Biotechnology Journal}, volume = {16}, journal = {Plant Biotechnology Journal}, number = {4}, publisher = {Wiley}, address = {Hoboken}, issn = {1467-7644}, doi = {10.1111/pbi.12840}, pages = {939 -- 950}, year = {2017}, abstract = {Potato (Solanum tuberosum L.) is one of the most important food crops worldwide. Current potato varieties are highly susceptible to drought stress. In view of global climate change, selection of cultivars with improved drought tolerance and high yield potential is of paramount importance. Drought tolerance breeding of potato is currently based on direct selection according to yield and phenotypic traits and requires multiple trials under drought conditions. Marker-assisted selection (MAS) is cheaper, faster and reduces classification errors caused by noncontrolled environmental effects. We analysed 31 potato cultivars grown under optimal and reduced water supply in six independent field trials. Drought tolerance was determined as tuber starch yield. Leaf samples from young plants were screened for preselected transcript and nontargeted metabolite abundance using qRT-PCR and GC-MS profiling, respectively. Transcript marker candidates were selected from a published RNA-Seq data set. A Random Forest machine learning approach extracted metabolite and transcript markers for drought tolerance prediction with low error rates of 6\% and 9\%, respectively. Moreover, by combining transcript and metabolite markers, the prediction error was reduced to 4.3\%. Feature selection from Random Forest models allowed model minimization, yielding a minimal combination of only 20 metabolite and transcript markers that were successfully tested for their reproducibility in 16 independent agronomic field trials. We demonstrate that a minimum combination of transcript and metabolite markers sampled at early cultivation stages predicts potato yield stability under drought largely independent of seasonal and regional agronomic conditions.}, language = {en} } @misc{SprengerErbanSeddigetal.2018, author = {Sprenger, Heike and Erban, Alexander and Seddig, Sylvia and Rudack, Katharina and Thalhammer, Anja and Le, Mai Q. and Walther, Dirk and Zuther, Ellen and K{\"o}hl, Karin I. and Kopka, Joachim and Hincha, Dirk K.}, title = {Metabolite and transcript markers for the prediction of potato drought tolerance}, series = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, journal = {Postprints der Universit{\"a}t Potsdam : Mathematisch-Naturwissenschaftliche Reihe}, number = {673}, issn = {1866-8372}, doi = {10.25932/publishup-42463}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-424630}, pages = {12}, year = {2018}, abstract = {Potato (Solanum tuberosum L.) is one of the most important food crops worldwide. Current potato varieties are highly susceptible to drought stress. In view of global climate change, selection of cultivars with improved drought tolerance and high yield potential is of paramount importance. Drought tolerance breeding of potato is currently based on direct selection according to yield and phenotypic traits and requires multiple trials under drought conditions. Marker-assisted selection (MAS) is cheaper, faster and reduces classification errors caused by noncontrolled environmental effects. We analysed 31 potato cultivars grown under optimal and reduced water supply in six independent field trials. Drought tolerance was determined as tuber starch yield. Leaf samples from young plants were screened for preselected transcript and nontargeted metabolite abundance using qRT-PCR and GC-MS profiling, respectively. Transcript marker candidates were selected from a published RNA-Seq data set. A Random Forest machine learning approach extracted metabolite and transcript markers for drought tolerance prediction with low error rates of 6\% and 9\%, respectively. Moreover, by combining transcript and metabolite markers, the prediction error was reduced to 4.3\%. Feature selection from Random Forest models allowed model minimization, yielding a minimal combination of only 20 metabolite and transcript markers that were successfully tested for their reproducibility in 16 independent agronomic field trials. We demonstrate that a minimum combination of transcript and metabolite markers sampled at early cultivation stages predicts potato yield stability under drought largely independent of seasonal and regional agronomic conditions.}, language = {en} } @article{SteinbergVasyuraBathkeGaebleretal.2021, author = {Steinberg, Andreas and Vasyura-Bathke, Hannes and Gaebler, Peter Jost and Ohrnberger, Matthias and Ceranna, Lars}, title = {Estimation of seismic moment tensors using variational inference machine learning}, series = {Journal of geophysical research : Solid earth}, volume = {126}, journal = {Journal of geophysical research : Solid earth}, number = {10}, publisher = {American Geophysical Union}, address = {Washington}, issn = {2169-9313}, doi = {10.1029/2021JB022685}, pages = {16}, year = {2021}, abstract = {We present an approach for rapidly estimating full moment tensors of earthquakes and their parameter uncertainties based on short time windows of recorded seismic waveform data by considering deep learning of Bayesian Neural Networks (BNNs). The individual neural networks are trained on synthetic seismic waveform data and corresponding known earthquake moment-tensor parameters. A monitoring volume has been predefined to form a three-dimensional grid of locations and to train a BNN for each grid point. Variational inference on several of these networks allows us to consider several sources of error and how they affect the estimated full moment-tensor parameters and their uncertainties. In particular, we demonstrate how estimated parameter distributions are affected by uncertainties in the earthquake centroid location in space and time as well as in the assumed Earth structure model. We apply our approach as a proof of concept on seismic waveform recordings of aftershocks of the Ridgecrest 2019 earthquake with moment magnitudes ranging from Mw 2.7 to Mw 5.5. Overall, good agreement has been achieved between inferred parameter ensembles and independently estimated parameters using classical methods. Our developed approach is fast and robust, and therefore, suitable for down-stream analyses that need rapid estimates of the source mechanism for a large number of earthquakes.}, language = {en} } @article{TongNikoloski2020, author = {Tong, Hao and Nikoloski, Zoran}, title = {Machine learning approaches for crop improvement}, series = {Journal of plant physiology : biochemistry, physiology, molecular biology and biotechnology of plants}, volume = {257}, journal = {Journal of plant physiology : biochemistry, physiology, molecular biology and biotechnology of plants}, publisher = {Elsevier}, address = {M{\"u}nchen}, issn = {0176-1617}, doi = {10.1016/j.jplph.2020.153354}, pages = {13}, year = {2020}, abstract = {Highly efficient and accurate selection of elite genotypes can lead to dramatic shortening of the breeding cycle in major crops relevant for sustaining present demands for food, feed, and fuel. In contrast to classical approaches that emphasize the need for resource-intensive phenotyping at all stages of artificial selection, genomic selection dramatically reduces the need for phenotyping. Genomic selection relies on advances in machine learning and the availability of genotyping data to predict agronomically relevant phenotypic traits. Here we provide a systematic review of machine learning approaches applied for genomic selection of single and multiple traits in major crops in the past decade. We emphasize the need to gather data on intermediate phenotypes, e.g. metabolite, protein, and gene expression levels, along with developments of modeling techniques that can lead to further improvements of genomic selection. In addition, we provide a critical view of factors that affect genomic selection, with attention to transferability of models between different environments. Finally, we highlight the future aspects of integrating high-throughput molecular phenotypic data from omics technologies with biological networks for crop improvement.}, language = {en} } @article{VaidChanChaudharyetal.2021, author = {Vaid, Akhil and Chan, Lili and Chaudhary, Kumardeep and Jaladanki, Suraj K. and Paranjpe, Ishan and Russak, Adam J. and Kia, Arash and Timsina, Prem and Levin, Matthew A. and He, John Cijiang and B{\"o}ttinger, Erwin and Charney, Alexander W. and Fayad, Zahi A. and Coca, Steven G. and Glicksberg, Benjamin S. and Nadkarni, Girish N.}, title = {Predictive approaches for acute dialysis requirement and death in COVID-19}, series = {Clinical journal of the American Society of Nephrology : CJASN}, volume = {16}, journal = {Clinical journal of the American Society of Nephrology : CJASN}, number = {8}, publisher = {American Society of Nephrology}, address = {Washington}, organization = {MSCIC}, issn = {1555-9041}, doi = {10.2215/CJN.17311120}, pages = {1158 -- 1168}, year = {2021}, abstract = {Background and objectives AKI treated with dialysis initiation is a common complication of coronavirus disease 2019 (COVID-19) among hospitalized patients. However, dialysis supplies and personnel are often limited. Design, setting, participants, \& measurements Using data from adult patients hospitalized with COVID-19 from five hospitals from theMount Sinai Health System who were admitted between March 10 and December 26, 2020, we developed and validated several models (logistic regression, Least Absolute Shrinkage and Selection Operator (LASSO), random forest, and eXtreme GradientBoosting [XGBoost; with and without imputation]) for predicting treatment with dialysis or death at various time horizons (1, 3, 5, and 7 days) after hospital admission. Patients admitted to theMount Sinai Hospital were used for internal validation, whereas the other hospitals formed part of the external validation cohort. Features included demographics, comorbidities, and laboratory and vital signs within 12 hours of hospital admission. Results A total of 6093 patients (2442 in training and 3651 in external validation) were included in the final cohort. Of the different modeling approaches used, XGBoost without imputation had the highest area under the receiver operating characteristic (AUROC) curve on internal validation (range of 0.93-0.98) and area under the precisionrecall curve (AUPRC; range of 0.78-0.82) for all time points. XGBoost without imputation also had the highest test parameters on external validation (AUROC range of 0.85-0.87, and AUPRC range of 0.27-0.54) across all time windows. XGBoost without imputation outperformed all models with higher precision and recall (mean difference in AUROC of 0.04; mean difference in AUPRC of 0.15). Features of creatinine, BUN, and red cell distribution width were major drivers of the model's prediction. Conclusions An XGBoost model without imputation for prediction of a composite outcome of either death or dialysis in patients positive for COVID-19 had the best performance, as compared with standard and other machine learning models.}, language = {en} } @article{VaidSomaniRussaketal.2020, author = {Vaid, Akhil and Somani, Sulaiman and Russak, Adam J. and De Freitas, Jessica K. and Chaudhry, Fayzan F. and Paranjpe, Ishan and Johnson, Kipp W. and Lee, Samuel J. and Miotto, Riccardo and Richter, Felix and Zhao, Shan and Beckmann, Noam D. and Naik, Nidhi and Kia, Arash and Timsina, Prem and Lala, Anuradha and Paranjpe, Manish and Golden, Eddye and Danieletto, Matteo and Singh, Manbir and Meyer, Dara and O'Reilly, Paul F. and Huckins, Laura and Kovatch, Patricia and Finkelstein, Joseph and Freeman, Robert M. and Argulian, Edgar and Kasarskis, Andrew and Percha, Bethany and Aberg, Judith A. and Bagiella, Emilia and Horowitz, Carol R. and Murphy, Barbara and Nestler, Eric J. and Schadt, Eric E. and Cho, Judy H. and Cordon-Cardo, Carlos and Fuster, Valentin and Charney, Dennis S. and Reich, David L. and B{\"o}ttinger, Erwin and Levin, Matthew A. and Narula, Jagat and Fayad, Zahi A. and Just, Allan C. and Charney, Alexander W. and Nadkarni, Girish N. and Glicksberg, Benjamin S.}, title = {Machine learning to predict mortality and critical events in a cohort of patients with COVID-19 in New York City: model development and validation}, series = {Journal of medical internet research : international scientific journal for medical research, information and communication on the internet ; JMIR}, volume = {22}, journal = {Journal of medical internet research : international scientific journal for medical research, information and communication on the internet ; JMIR}, number = {11}, publisher = {Healthcare World}, address = {Richmond, Va.}, issn = {1439-4456}, doi = {10.2196/24018}, pages = {19}, year = {2020}, abstract = {Background: COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking. Objective: The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points. Methods: We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions. Results: Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction. Conclusions: We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.}, language = {en} } @book{Weber2023, author = {Weber, Benedikt}, title = {Human pose estimation for decubitus prophylaxis}, number = {153}, publisher = {Universit{\"a}tsverlag Potsdam}, address = {Potsdam}, isbn = {978-3-86956-551-4}, issn = {1613-5652}, doi = {10.25932/publishup-56719}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-567196}, publisher = {Universit{\"a}t Potsdam}, pages = {73}, year = {2023}, abstract = {Decubitus is one of the most relevant diseases in nursing and the most expensive to treat. It is caused by sustained pressure on tissue, so it particularly affects bed-bound patients. This work lays a foundation for pressure mattress-based decubitus prophylaxis by implementing a solution to the single-frame 2D Human Pose Estimation problem. For this, methods of Deep Learning are employed. Two approaches are examined, a coarse-to-fine Convolutional Neural Network for direct regression of joint coordinates and a U-Net for the derivation of probability distribution heatmaps. We conclude that training our models on a combined dataset of the publicly available Bodies at Rest and SLP data yields the best results. Furthermore, various preprocessing techniques are investigated, and a hyperparameter optimization is performed to discover an improved model architecture. Another finding indicates that the heatmap-based approach outperforms direct regression. This model achieves a mean per-joint position error of 9.11 cm for the Bodies at Rest data and 7.43 cm for the SLP data. We find that it generalizes well on data from mattresses other than those seen during training but has difficulties detecting the arms correctly. Additionally, we give a brief overview of the medical data annotation tool annoto we developed in the bachelor project and furthermore conclude that the Scrum framework and agile practices enhanced our development workflow.}, language = {en} } @article{WilkschAbramova2023, author = {Wilksch, Moritz and Abramova, Olga}, title = {PyFin-sentiment}, series = {International journal of information management data insights}, volume = {3}, journal = {International journal of information management data insights}, number = {1}, publisher = {Elsevier}, address = {Amsterdam}, issn = {2667-0968}, doi = {10.1016/j.jjimei.2023.100171}, pages = {10}, year = {2023}, abstract = {Responding to the poor performance of generic automated sentiment analysis solutions on domain-specific texts, we collect a dataset of 10,000 tweets discussing the topics of finance and investing. We manually assign each tweet its market sentiment, i.e., the investor's anticipation of a stock's future return. Using this data, we show that all existing sentiment models trained on adjacent domains struggle with accurate market sentiment analysis due to the task's specialized vocabulary. Consequently, we design, train, and deploy our own sentiment model. It outperforms all previous models (VADER, NTUSD-Fin, FinBERT, TwitterRoBERTa) when evaluated on Twitter posts. On posts from a different platform, our model performs on par with BERT-based large language models. We achieve this result at a fraction of the training and inference costs due to the model's simple design. We publish the artifact as a python library to facilitate its use by future researchers and practitioners.}, language = {en} } @article{WulffBuschhueterWestphaletal.2020, author = {Wulff, Peter and Buschh{\"u}ter, David and Westphal, Andrea and Nowak, Anna and Becker, Lisa and Robalino, Hugo and Stede, Manfred and Borowski, Andreas}, title = {Computer-based classification of preservice physics teachers' written reflections}, series = {Journal of science education and technology}, volume = {30}, journal = {Journal of science education and technology}, number = {1}, publisher = {Springer}, address = {Dordrecht}, issn = {1059-0145}, doi = {10.1007/s10956-020-09865-1}, pages = {1 -- 15}, year = {2020}, abstract = {Reflecting in written form on one's teaching enactments has been considered a facilitator for teachers' professional growth in university-based preservice teacher education. Writing a structured reflection can be facilitated through external feedback. However, researchers noted that feedback in preservice teacher education often relies on holistic, rather than more content-based, analytic feedback because educators oftentimes lack resources (e.g., time) to provide more analytic feedback. To overcome this impediment to feedback for written reflection, advances in computer technology can be of use. Hence, this study sought to utilize techniques of natural language processing and machine learning to train a computer-based classifier that classifies preservice physics teachers' written reflections on their teaching enactments in a German university teacher education program. To do so, a reflection model was adapted to physics education. It was then tested to what extent the computer-based classifier could accurately classify the elements of the reflection model in segments of preservice physics teachers' written reflections. Multinomial logistic regression using word count as a predictor was found to yield acceptable average human-computer agreement (F1-score on held-out test dataset of 0.56) so that it might fuel further development towards an automated feedback tool that supplements existing holistic feedback for written reflections with data-based, analytic feedback.}, language = {en} } @article{WulffMientusNowaketal.2023, author = {Wulff, Peter and Mientus, Lukas and Nowak, Anna and Borowski, Andreas}, title = {KI-basierte Auswertung von schriftlichen Unterrichtsreflexionen im Fach Physik und automatisierte R{\"u}ckmeldung}, series = {PSI-Potsdam: Ergebnisbericht zu den Aktivit{\"a}ten im Rahmen der Qualit{\"a}tsoffensive Lehrerbildung (2019-2023) (Potsdamer Beitr{\"a}ge zur Lehrerbildung und Bildungsforschung ; 3)}, journal = {PSI-Potsdam: Ergebnisbericht zu den Aktivit{\"a}ten im Rahmen der Qualit{\"a}tsoffensive Lehrerbildung (2019-2023) (Potsdamer Beitr{\"a}ge zur Lehrerbildung und Bildungsforschung ; 3)}, number = {3}, publisher = {Universit{\"a}tsverlag Potsdam}, address = {Potsdam}, isbn = {978-3-86956-568-2}, issn = {2626-3556}, doi = {10.25932/publishup-61636}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-616363}, pages = {103 -- 115}, year = {2023}, abstract = {F{\"u}r die Entwicklung professioneller Handlungskompetenzen angehender Lehrkr{\"a}fte stellt die Unterrichtsreflexion ein wichtiges Instrument dar, um Theoriewissen und Praxiserfahrungen in Beziehung zu setzen. Die Auswertung von Unterrichtsreflexionen und eine entsprechende R{\"u}ckmeldung stellt Forschende und Dozierende allerdings vor praktische wie theoretische Herausforderungen. Im Kontext der Forschung zu K{\"u}nstlicher Intelligenz (KI) entwickelte Methoden bieten hier neue Potenziale. Der Beitrag stellt {\"u}berblicksartig zwei Teilstudien vor, die mit Hilfe von KI-Methoden wie dem maschinellen Lernen untersuchen, inwieweit eine Auswertung von Unterrichtsreflexionen angehender Physiklehrkr{\"a}fte auf Basis eines theoretisch abgeleiteten Reflexionsmodells und die automatisierte R{\"u}ckmeldung hierzu m{\"o}glich sind. Dabei wurden unterschiedliche Ans{\"a}tze des maschinellen Lernens verwendet, um modellbasierte Klassifikation und Exploration von Themen in Unterrichtsreflexionen umzusetzen. Die Genauigkeit der Ergebnisse wurde vor allem durch sog. Große Sprachmodelle gesteigert, die auch den Transfer auf andere Standorte und F{\"a}cher erm{\"o}glichen. F{\"u}r die fachdidaktische Forschung bedeuten sie jedoch wiederum neue Herausforderungen, wie etwa systematische Verzerrungen und Intransparenz von Entscheidungen. Dennoch empfehlen wir, die Potenziale der KI-basierten Methoden gr{\"u}ndlicher zu erforschen und konsequent in der Praxis (etwa in Form von Webanwendungen) zu implementieren.}, language = {de} } @phdthesis{Zali2023, author = {Zali, Zahra}, title = {Volcanic tremor analysis based on advanced signal processing concepts including music information retrieval (MIR) strategies}, doi = {10.25932/publishup-61086}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-610866}, school = {Universit{\"a}t Potsdam}, pages = {viii, 95}, year = {2023}, abstract = {Volcanoes are one of the Earth's most dynamic zones and responsible for many changes in our planet. Volcano seismology aims to provide an understanding of the physical processes in volcanic systems and anticipate the style and timing of eruptions by analyzing the seismic records. Volcanic tremor signals are usually observed in the seismic records before or during volcanic eruptions. Their analysis contributes to evaluate the evolving volcanic activity and potentially predict eruptions. Years of continuous seismic monitoring now provide useful information for operational eruption forecasting. The continuously growing amount of seismic recordings, however, poses a challenge for analysis, information extraction, and interpretation, to support timely decision making during volcanic crises. Furthermore, the complexity of eruption processes and precursory activities makes the analysis challenging. A challenge in studying seismic signals of volcanic origin is the coexistence of transient signal swarms and long-lasting volcanic tremor signals. Separating transient events from volcanic tremors can, therefore, contribute to improving our understanding of the underlying physical processes. Some similar issues (data reduction, source separation, extraction, and classification) are addressed in the context of music information retrieval (MIR). The signal characteristics of acoustic and seismic recordings comprise a number of similarities. This thesis is going beyond classical signal analysis techniques usually employed in seismology by exploiting similarities of seismic and acoustic signals and building the information retrieval strategy on the expertise developed in the field of MIR. First, inspired by the idea of harmonic-percussive separation (HPS) in musical signal processing, I have developed a method to extract harmonic volcanic tremor signals and to detect transient events from seismic recordings. This provides a clean tremor signal suitable for tremor investigation along with a characteristic function suitable for earthquake detection. Second, using HPS algorithms, I have developed a noise reduction technique for seismic signals. This method is especially useful for denoising ocean bottom seismometers, which are highly contaminated by noise. The advantage of this method compared to other denoising techniques is that it doesn't introduce distortion to the broadband earthquake waveforms, which makes it reliable for different applications in passive seismological analysis. Third, to address the challenge of extracting information from high-dimensional data and investigating the complex eruptive phases, I have developed an advanced machine learning model that results in a comprehensive signal processing scheme for volcanic tremors. Using this method seismic signatures of major eruptive phases can be automatically detected. This helps to provide a chronology of the volcanic system. Also, this model is capable to detect weak precursory volcanic tremors prior to the eruption, which could be used as an indicator of imminent eruptive activity. The extracted patterns of seismicity and their temporal variations finally provide an explanation for the transition mechanism between eruptive phases.}, language = {en} } @book{ZhangPlauthEberhardtetal.2020, author = {Zhang, Shuhao and Plauth, Max and Eberhardt, Felix and Polze, Andreas and Lehmann, Jens and Sejdiu, Gezim and Jabeen, Hajira and Servadei, Lorenzo and M{\"o}stl, Christian and B{\"a}r, Florian and Netzeband, Andr{\´e} and Schmidt, Rainer and Knigge, Marlene and Hecht, Sonja and Prifti, Loina and Krcmar, Helmut and Sapegin, Andrey and Jaeger, David and Cheng, Feng and Meinel, Christoph and Friedrich, Tobias and Rothenberger, Ralf and Sutton, Andrew M. and Sidorova, Julia A. and Lundberg, Lars and Rosander, Oliver and Sk{\"o}ld, Lars and Di Varano, Igor and van der Walt, Est{\´e}e and Eloff, Jan H. P. and Fabian, Benjamin and Baumann, Annika and Ermakova, Tatiana and Kelkel, Stefan and Choudhary, Yash and Cooray, Thilini and Rodr{\´i}guez, Jorge and Medina-P{\´e}rez, Miguel Angel and Trejo, Luis A. and Barrera-Animas, Ari Yair and Monroy-Borja, Ra{\´u}l and L{\´o}pez-Cuevas, Armando and Ram{\´i}rez-M{\´a}rquez, Jos{\´e} Emmanuel and Grohmann, Maria and Niederleithinger, Ernst and Podapati, Sasidhar and Schmidt, Christopher and Huegle, Johannes and de Oliveira, Roberto C. L. and Soares, F{\´a}bio Mendes and van Hoorn, Andr{\´e} and Neumer, Tamas and Willnecker, Felix and Wilhelm, Mathias and Kuster, Bernhard}, title = {HPI Future SOC Lab - Proceedings 2017}, number = {130}, editor = {Meinel, Christoph and Polze, Andreas and Beins, Karsten and Strotmann, Rolf and Seibold, Ulrich and R{\"o}dszus, Kurt and M{\"u}ller, J{\"u}rgen}, publisher = {Universit{\"a}tsverlag Potsdam}, address = {Potsdam}, isbn = {978-3-86956-475-3}, issn = {1613-5652}, doi = {10.25932/publishup-43310}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-433100}, publisher = {Universit{\"a}t Potsdam}, pages = {ix, 235}, year = {2020}, abstract = {The "HPI Future SOC Lab" is a cooperation of the Hasso Plattner Institute (HPI) and industry partners. Its mission is to enable and promote exchange and interaction between the research community and the industry partners. The HPI Future SOC Lab provides researchers with free of charge access to a complete infrastructure of state of the art hard and software. This infrastructure includes components, which might be too expensive for an ordinary research environment, such as servers with up to 64 cores and 2 TB main memory. The offerings address researchers particularly from but not limited to the areas of computer science and business information systems. Main areas of research include cloud computing, parallelization, and In-Memory technologies. This technical report presents results of research projects executed in 2017. Selected projects have presented their results on April 25th and November 15th 2017 at the Future SOC Lab Day events.}, language = {en} } @inproceedings{OPUS4-40678, title = {HPI Future SOC Lab}, editor = {Meinel, Christoph and Polze, Andreas and Oswald, Gerhard and Strotmann, Rolf and Seibold, Ulrich and Schulzki, Bernhard}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-406787}, pages = {iii, 180}, year = {2016}, abstract = {The "HPI Future SOC Lab" is a cooperation of the Hasso Plattner Institute (HPI) and industrial partners. Its mission is to enable and promote exchange and interaction between the research community and the industrial partners. The HPI Future SOC Lab provides researchers with free of charge access to a complete infrastructure of state of the art hard and software. This infrastructure includes components, which might be too expensive for an ordinary research environment, such as servers with up to 64 cores and 2 TB main memory. The offerings address researchers particularly from but not limited to the areas of computer science and business information systems. Main areas of research include cloud computing, parallelization, and In-Memory technologies. This technical report presents results of research projects executed in 2016. Selected projects have presented their results on April 5th and November 3th 2016 at the Future SOC Lab Day events.}, language = {en} }